Benchmark Almanac

Angry Tweets

GPT-6.2 Astra 67% leads since November 2026
Not in model cards

Models classify the sentiment of Danish social media posts using labels supplied by crowdworkers.

EuroEval scores sentiment labels with the Matthews correlation coefficient.

Released
May 2021
Kind
Reading in many languages
Level
Everyday
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Language
Danish
In and out
Text → Text
Interaction
Single turn
Refresh
One fixed set
Run by
Anyone
Strongest model tested
GPT-6.2 Astra #1 on the capability index
Last model added
November 2026
Scores refreshed
November 20th, 2026
596
Entries on its own leaderboard, 149 models we track

GPT-6.2 Astra leads with 67% GPT-6.2 Astra leads with 67%

Best score per entry from the benchmark's own leaderboard, read on November 20th, 2026. Where a model ran with several settings or harnesses, we keep its best. Every run is in the next chart. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. This view shows only models we track, each at its best.

596 entries. Scroll sideways for the rest.

  1. 67.0 GPT-6.2 Astra
  2. 66.3 GLM-5.3-Flash
  3. 65.3 GPT-6 Astra
  4. 64.5 mistralai/Mistral-Large-3-675B-Instruct-2512
  5. 64.3 claude-sonnet-4-5-20250929#thinking
  6. 64.2 Gemini 3.7 Flash
  7. 64.0 danish-foundation-models/DFM-Mimir-v1.5
  8. 63.8 Gemini 2.5 Flash
  9. 63.6 speakleash/Bielik-11B-v3.1-Instruct
  10. 63.5 Qwen/Qwen3.5-27B-FP8
  11. 63.5 Gemma 4 31B IT
  12. 63.3 NbAiLab/borealis2-31b-preview
  13. 63.2 claude-3-7-sonnet-20250219#no-thinking
  14. 63.2 Qwen/Qwen3.6-27B-FP8
  15. 63.2 syvai/danskgpt-v4-31b
  16. 63.2 deepseek/deepseek-flash#no-thinking
  17. 62.8 GPT-6.2 Sol
  18. 62.6 speakleash/Bielik-11B-v3.0-Instruct
  19. 62.6 claude-sonnet-4-5-20250929#no-thinking
  20. 62.6 GPT-6.1 Sol
  21. 62.2 Mistral Medium 3.5
  22. 62.2 o3-mini
  23. 61.9 Gemini 3.5 Flash
  24. 61.7 gemini/gemini-3-flash-preview#thinking
  25. 61.6 zai-org/GLM-4.7-FP8
  26. 61.5 danish-foundation-models/DFM-Mimir-v1.8
  27. 60.9 Qwen3.5 397B-A17B
  28. 60.8 meta-llama/Llama-3.1-405B-Instruct-FP8
  29. 60.6 Claude Opus 4.8
  30. 60.6 deepseek/deepseek-flash#low
  31. 60.6 Qwen 3.8 27B
  32. 60.5 Qwen/Qwen3.5-35B-A3B-FP8
  33. 60.2 GPT-6 Luna
  34. 60.2 Gemini 3.6 Flash
  35. 60.2 Gemini 3 Pro
  36. 60.0 Olmo 3.1 32B Think
  37. 60.0 Gemma 4 12B
  38. 59.8 NbAiLab/borealis-open-27b
  39. 59.6 danish-foundation-models/DFM-Mimir
  40. 59.5 NbAiLab/borealis-27b-instruct-preview
  41. 59.3 GPT-5.6 Sol
  42. 59.3 Qwen/Qwen3-235B-A22B-Thinking-2507-FP8
  43. 59.1 LumiOpen/Llama-Poro-2-70B-SFT
  44. 59.1 mistralai/Mistral-Small-3.1-24B-Instruct-2503
  45. 58.9 ThatsGroes/gemma-2-27b-it-FP8-Dynamic
  46. 58.8 GPT-5
  47. 58.8 deepcogito/cogito-v2-preview-llama-70B
  48. 58.7 gemini/gemini-3-flash-preview#no-thinking
  49. 58.7 Reka Flash 3.1
  50. 58.6 NbAiLab/borealis-27b
  51. 58.6 Cogito v1 Preview 70B
  52. 58.5 LumiOpen/Llama-Poro-2-70B-Instruct
  53. 58.5 GPT-5.2
  54. 58.5 glm-4.7-flash
  55. 58.4 Llama 3.1-70B
  56. 58.3 claude-3-7-sonnet-20250219#thinking
  57. 58.3 mistralai/Mistral-Small-4-119B-2603
  58. 58.2 Qwen3.5-9B
  59. 58.2 GadflyII/GLM-4.7-Flash-NVFP4
  60. 58.1 openai/gpt-5.4-mini-2026-03-17#medium
  61. 58.1 Gemini 3.1 Flash-Lite
  62. 58.1 o3
  63. 58.1 Llama 4 Scout
  64. 58.0 Reka Flash 3
  65. 58.0 EXAONE 4.0 32B
  66. 57.9 syvai/danskgpt-v4-12b
  67. 57.7 Claude Sonnet 4.6
  68. 57.6 google/gemma-2-27b-it
  69. 57.5 o4-mini
  70. 57.4 Gemma 3 27B
  71. 57.3 mistralai/Mistral-Small-24B-Instruct-2501
  72. 57.1 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16
  73. 57.0 GPT-5.5
  74. 57.0 speakleash/Bielik-11B-v2.3-Instruct
  75. 57.0 typesafe/jev-1.13.0
  76. 56.9 Gemma 4 26B A4B
  77. 56.9 GPT-4.1
  78. 56.9 Nemotron-Cascade 2 30B-A3B
  79. 56.8 Llama 3.3 70B Instruct
  80. 56.8 Unbabel/Tower-Plus-72B
  81. 56.8 mistralai/Mistral-Small-3.2-24B-Instruct-2506
  82. 56.7 gpt-5-nano-2025-08-07#high
  83. 56.7 mistralai/Mistral-Small-3.1-24B-Base-2503
  84. 56.7 Gemma 3 12B
  85. 56.6 QwQ-32B
  86. 56.4 ollama_chat/gemma3:27b-it-fp16
  87. 56.4 Apertus 1.5 70B
  88. 56.3 NbAiLab/borealis-12b-instruct-preview
  89. 56.2 OLMo 2 32B Instruct
  90. 56.2 danish-foundation-models/munin-qwen3.5-9B
  91. 56.2 GLM-4.5-Air
  92. 56.2 gpt-5-2025-08-07#high
  93. 56.2 ollama_chat/gemma3:12b-it-q4_K_M
  94. 56.1 deepcogito/cogito-v1-preview-qwen-32B
  95. 56.1 Qwen3.5-35B-A3B-Base
  96. 56.1 NbAiLab/borealis2-26b-a4b-preview
  97. 56.0 ollama_chat/gemma3:27b-it-q4_K_M
  98. 56.0 Aya Expanse 8B
  99. 56.0 GPT-5 nano
  100. 56.0 mistralai/Magistral-Small-2509
  101. 55.9 Qwen/Qwen3-235B-A22B-Instruct-2507-FP8
  102. 55.9 Claude Haiku 4.5
  103. 55.9 Grok 4.20
  104. 55.7 openai/gpt-5.4-2026-03-05#none
  105. 55.7 NovaSky-AI/Sky-T1-32B-Preview
  106. 55.7 Qwen/Qwen3.5-9B#no-thinking
  107. 55.7 Gemini 2.5 Pro
  108. 55.6 zai-org/GLM-4.5-Air#no-thinking
  109. 55.6 gpt-oss-120b
  110. 55.5 Qwen3-14B
  111. 55.5 NbAiLab/borealis2-e4b-preview
  112. 55.5 Ministral 3 3B Instruct
  113. 55.5 GPT-5.6 Terra
  114. 55.4 Apertus 70B
  115. 55.4 Gemma 3n E4B
  116. 55.4 Qwen3-235B-A22B
  117. 55.3 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8
  118. 55.2 Qwen/Qwen3-14B#no-thinking
  119. 55.2 Aleph-Alpha/Kolibri-1
  120. 55.2 meta-llama/Llama-3.1-8B-Instruct
  121. 55.2 Gemma 3n E2B
  122. 55.2 GPT-6 Sol
  123. 55.1 Gemini 3.5 Flash-Lite
  124. 55.1 Qwen3-Next-80B-A3B
  125. 55.1 openai/gpt-5.4-mini-2026-03-17#high
  126. 55.0 NbAiLab/borealis-open-12b
  127. 55.0 Grok-2
  128. 55.0 Phi-4
  129. 54.9 deepcogito/cogito-v1-preview-qwen-14B
  130. 54.8 Grok 3
  131. 54.7 speakleash/Bielik-Minitron-7B-v3.0-Instruct
  132. 54.6 Qwen/Qwen3-32B#no-thinking
  133. 54.6 Voxtral Small
  134. 54.5 Qwen3-8B
  135. 54.5 gpt-oss-20b
  136. 54.5 Grok-3 mini
  137. 54.4 NousResearch/DeepHermes-3-Llama-3-8B-Preview
  138. 54.4 utter-project/EuroLLM-9B-Instruct
  139. 54.3 Gemma 3 27B PT
  140. 54.2 deepcogito/cogito-v1-preview-llama-8B
  141. 54.2 NbAiLab/borealis-12b
  142. 54.1 timpal0l/BeagleCatMunin
  143. 54.1 Grok 4.1 Fast
  144. 54.1 Aya Expanse 32B
  145. 54.0 alexandrainst/Meta-Llama-3.1-8B-laerebogen-and-dialog-k2-da
  146. 54.0 GLM-4.1V-9B-Thinking
  147. 53.9 openai/gpt-5.4-mini-2026-03-17#low
  148. 53.6 ollama_chat/gemma3:12b-it-fp16
  149. 53.6 gpt-5-2025-08-07#minimal
  150. 53.4 alexandrainst/Meta-Llama-3.1-8B-laerebogen
  151. 53.4 gemini/gemini-2.5-flash#no-thinking
  152. 53.3 Qwen3-Next-80B-A3B-Instruct
  153. 53.3 ollama_chat/phi4:14b-q4_K_M
  154. 53.2 allenai/Olmo-3.1-32B-Instruct-DPO
  155. 53.2 KennethEnevoldsen/dfm-sentence-encoder-large
  156. 53.2 gemini/gemini-2.5-flash-lite-preview-06-17#thinking
  157. 53.2 Gemini 2.5 Flash-Lite
  158. 53.2 google/gemma-2-9b-it
  159. 53.1 utter-project/EuroLLM-22B-Instruct-2512
  160. 53.0 gpt-5-mini-2025-08-07#high
  161. 53.0 Gemma 3 12B PT
  162. 53.0 utter-project/EuroLLM-9B
  163. 52.9 allenai/Llama-3.1-Tulu-3-8B-DPO
  164. 52.9 intfloat/multilingual-e5-large
  165. 52.9 mistralai/Ministral-3-14B-Base-2512
  166. 52.8 allenai/Olmo-3.1-32B-Instruct-SFT
  167. 52.8 Ministral 3 14B
  168. 52.8 danish-foundation-models/munin-ministral3-8B
  169. 52.8 DeepSeek-R1-Distill-Llama-70B
  170. 52.8 Gemma 4 E4B
  171. 52.7 GPT-5 mini
  172. 52.7 ordbogen/odin-medium
  173. 52.7 Qwen/Qwen3-30B-A3B#no-thinking
  174. 52.6 timpal0l/dolphin-2.9-llama3-8b-flashback
  175. 52.6 Llama-3.1-Tulu-3-8B
  176. 52.5 GPT-5.6 Luna
  177. 52.5 merge-crew/da-sv-ties
  178. 52.5 utter-project/EuroLLM-22B-Instruct-Preview
  179. 52.3 Olmo 3.1 32B Instruct
  180. 52.3 alexandrainst/Meta-Llama-3.1-8B-laerebogen-two-epochs
  181. 52.3 01-ai/Yi-1.5-9B
  182. 52.2 birgermoell/BeagleCatMunin-Flashback-Bellman
  183. 52.2 mistralai/Ministral-3-14B-Instruct-2512
  184. 52.1 Command R7B
  185. 52.1 Qwen3-4B-Thinking-2507
  186. 52.1 ordbogen/odin-large
  187. 52.0 RJuro/munin-neuralbeagle-7b
  188. 52.0 meta-llama/Llama-4-Scout-17B-16E
  189. 52.0 speakleash/Bielik-11B-v2.6-Instruct
  190. 51.9 bineric/lynx-instruct-30b
  191. 51.9 Llama 3.1-405B
  192. 51.9 merge-crew/da-sv-dare-ties-density-0.9
  193. 51.8 openGPT-X/Teuken-7B-instruct-v0.6
  194. 51.8 merge-crew/da-sv-task-arithmetic
  195. 51.8 Granite 4.0 H Small
  196. 51.6 birgermoell/Flashback-Bellman
  197. 51.6 birgermoell/Rapid-Cycling
  198. 51.6 TrustLLMeu/baseline-7-8b_2-3t-tokens_llama
  199. 51.6 Qwen/Qwen3-8B#no-thinking
  200. 51.6 openai/gpt-5.4-mini-2026-03-17#none
  201. 51.5 merge-crew/da-sv-slerp
  202. 51.5 gemini/gemini-2.5-flash-lite-preview-06-17#no-thinking
  203. 51.5 gemini/gemini-2.5-flash-lite#no-thinking
  204. 51.3 CohereForAI/aya-23-35B
  205. 51.3 openai/gpt-5.4-nano-2026-03-17#medium
  206. 51.1 RJuro/munin-neuralbeagle-SkoleGPTOpenOrca-7b
  207. 51.1 LumiOpen/Viking-33B
  208. 51.1 bullerwins/Magistral-Small-2506-fp8
  209. 51.1 Trinity Mini
  210. 51.0 Qwen1.5-14B-Chat
  211. 50.9 danish-foundation-models/Meta-Llama-3.1-8B-laerebogen
  212. 50.9 tiiuae/falcon-11B
  213. 50.8 Granite-3.0-8B-Instruct
  214. 50.8 Qwen3.5-4B
  215. 50.7 NbAiLab/nb-roberta-base-scandi
  216. 50.7 openai/gpt-5.4-nano-2026-03-17#low
  217. 50.7 danish-foundation-models/DFM-Mimir-v1.8-Mini
  218. 50.6 Unbabel/Tower-Plus-9B
  219. 50.5 allenai/Llama-3.1-Tulu-3.1-8B
  220. 50.5 alexandrainst/Meta-Llama-3.1-8B-laerebogen-and-filtered-dialog-k2-da
  221. 50.5 meta-llama/Llama-2-70b-hf
  222. 50.5 ollama_chat/hf.co/google/gemma-3-4b-it-qat-q4_0-gguf
  223. 50.3 ollama_chat/gemma3:4b-it-q4_K_M
  224. 50.3 openai/gpt-5.4-nano-2026-03-17#none
  225. 50.3 senseable/WestLake-7B-v2
  226. 50.3 danish-foundation-models/encoder-large-v1
  227. 50.3 merge-crew/da-sv-dare-ties-density-0.6
  228. 50.2 152334H/miqu-1-70b-sf
  229. 50.1 mistralai/Ministral-3-8B-Reasoning-2512
  230. 50.0 deepcogito/cogito-v1-preview-llama-3B
  231. 50.0 KennethEnevoldsen/dfm-sentence-encoder-medium
  232. 49.8 intfloat/multilingual-e5-base
  233. 49.8 meta-llama/Llama-2-70b-chat-hf
  234. 49.7 NbAiLab/nb-roberta-base-scandi-1e4
  235. 49.7 AI-Sweden-Models/Llama-3-8B-instruct
  236. 49.7 NbAiLab/nb-roberta-base-ncc-plus-scandi-1e4
  237. 49.7 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-Base-BF16
  238. 49.6 AI-Sweden-Models/roberta-large-1160k
  239. 49.6 AI-Sweden-Models/tyr
  240. 49.6 openai/gpt-5.4-nano-2026-03-17#high
  241. 49.5 Gemma 4 E2B
  242. 49.5 NbAiLab/borealis-4b
  243. 49.5 utter-project/EuroLLM-22B-Preview
  244. 49.5 DeepSeek-R1-Distill-Qwen-32B
  245. 49.5 NbAiLab/nb-roberta-base-ncc-plus-scandi-2e4
  246. 49.4 Ministral 3 8B Instruct
  247. 49.3 OpenLLM-France/Luciole-23B-Instruct-1.1
  248. 49.3 Qwen/Qwen3-4B#no-thinking
  249. 49.3 Qwen3-4B
  250. 49.3 utter-project/EuroLLM-9B-Instruct-2512
  251. 49.3 LumiOpen/Viking-13B
  252. 49.2 Unbabel/Tower-Plus-2B
  253. 49.1 upstage/SOLAR-10.7B-v1.0
  254. 49.1 gpt-5-nano-2025-08-07#minimal
  255. 49.1 KennethEnevoldsen/munin_mistral-7b
  256. 49.1 Falcon3-10B-Instruct
  257. 49.0 deepseek-r1-0528-qwen3-8b
  258. 49.0 Gemma 2 27B
  259. 49.0 gpt-5-mini-2025-08-07#minimal
  260. 48.9 NbAiLab/borealis2-e2b-preview
  261. 48.8 timpal0l/BeagleCatMunin2
  262. 48.7 Ministral 3 3B Reasoning
  263. 48.7 amalia-llm/AMALIA-9B-0626-DPO
  264. 48.7 danish-foundation-models/munin-apertus-8b
  265. 48.6 synquid/gemma-3-4b-pt-laerebogen-with-follow-ups-lora
  266. 48.6 Tiny Aya Water
  267. 48.5 allenai/Olmo-3-7B-Think
  268. 48.4 meta-llama/Llama-2-13b-chat-hf
  269. 48.2 NbAiLab/borealis-4b-instruct-preview
  270. 48.2 NbAiLab/borealis-open-4b
  271. 48.1 Granite-3.0-8B-Base
  272. 48.1 CohereLabs/tiny-aya-base
  273. 48.1 vesteinn/DanskBERT
  274. 48.0 amalia-llm/AMALIA-9B-0626-SFT
  275. 48.0 Llama 3.1-8B
  276. 48.0 mistralai/Ministral-3-3B-Base-2512
  277. 48.0 nvidia-nemotron-nano-9b-v2
  278. 48.0 granite-4.0-micro
  279. 47.9 Tiny Aya Global
  280. 47.8 Falcon3-7B-Instruct
  281. 47.7 swiss-ai/Apertus-v1.1-4B
  282. 47.7 allenai/OLMo-2-1124-13B-Instruct
  283. 47.6 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16#no-thinking
  284. 47.6 mistralai/Ministral-3-8B-Base-2512
  285. 47.6 neph1/bellman-7b-mistral-instruct-v0.2
  286. 47.5 allenai/Llama-3.1-Tulu-3-8B-SFT
  287. 47.5 sentence-transformers/use-cmlm-multilingual
  288. 47.5 tiiuae/Falcon3-10B-Base
  289. 47.4 qwen3-4b-instruct-2507
  290. 47.4 Mistral 7B v0.3
  291. 47.4 mlabonne/NeuralBeagle14-7B
  292. 47.4 vesteinn/FoBERT
  293. 47.3 Gemma 7B
  294. 47.3 openeurollm/prelude-long
  295. 47.2 occiglot/occiglot-7b-eu5
  296. 47.2 bineric/lynx-instruct-30b#no-thinking
  297. 47.1 Gemma 2 2B IT
  298. 47.1 domyn/Domyn-Small-v1.0
  299. 46.9 DGurgurov/xlm-r_dan-latn
  300. 46.9 norallm/normistral-11b-long
  301. 46.8 bineric/lynx-instruct-30b#thinking
  302. 46.8 openeurollm/oellm-9b-128k-theta32m-v3
  303. 46.7 openeurollm/prelude-base
  304. 46.6 Yi-1.5-34B
  305. 46.3 mistralai/Ministral-8B-Instruct-2410
  306. 46.2 mii-llm/nesso-4B
  307. 46.2 swiss-ai/Apertus-v1.1-4B-Instruct
  308. 46.2 vesteinn/ScandiBERT-no-faroese
  309. 46.1 nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16
  310. 45.9 DeepSeek-R1-Distill-Qwen-14B
  311. 45.8 Apertus 1.5 8B
  312. 45.7 Qwen3-1.7B
  313. 45.6 allenai/OLMo-2-1124-7B-Instruct
  314. 45.5 Qwen/Qwen3-1.7B#no-thinking
  315. 45.2 Mistral 7B v0.1
  316. 45.1 alexandrainst/Meta-Llama-3.1-8B-dialog-k2-da
  317. 45.1 TrustLLMeu/baseline-7-8b_1t-tokens_llama
  318. 45.0 HPLT/NorOLMo-13B
  319. 44.9 Mistral-7B-Instruct-v0.2
  320. 44.7 Mistral-7B-Instruct-v0.1
  321. 44.6 LiquidAI/LFM2.5-8B-A1B-Base
  322. 44.6 tiiuae/Falcon3-7B-Base
  323. 44.5 birgermoell/Munin-NeuralBeagle-NorskGPT
  324. 44.5 ai21labs/Jamba-v0.1
  325. 44.5 RuterNorway/Llama-2-13b-chat-norwegian
  326. 44.4 norallm/normistral-11b-thinking
  327. 44.3 baidu/ERNIE-4.5-21B-A3B-PT
  328. 44.3 allenai/Olmo-3-1025-7B
  329. 44.2 mistralai/Voxtral-Mini-3B-2507
  330. 44.1 swiss-ai/Apertus-8B-2509
  331. 44.1 setu4993/LaBSE
  332. 44.0 AI-Sweden-Models/ModernBERT-large
  333. 44.0 ltg/norbert3-large
  334. 43.9 deepseek-ai/DeepSeek-R1-Distill-Llama-8B
  335. 43.8 nvidia/NVIDIA-Nemotron-3-Nano-4B-FP8
  336. 43.8 NbAiLab/nb-bert-ncc-male2female
  337. 43.7 NbAiLab/nb-roberta-base-ncc
  338. 43.4 Qwen1.5-14B
  339. 43.1 norallm/normistral-11b-warm
  340. 43.0 deahmed/Qwen3.5-2B-da-task
  341. 43.0 speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c
  342. 42.9 NbAiLab/nb-bert-base
  343. 42.7 Qwen3.5-2B
  344. 42.6 mlabonne/AlphaMonarch-7B
  345. 42.6 speakleash/Bielik-4.5B-v3.0-Instruct
  346. 42.5 occiglot/occiglot-7b-eu5-instruct
  347. 42.5 AFM-4.5B
  348. 42.3 ltg/norbert3-base
  349. 42.2 meta-llama/Llama-3.2-3B-Instruct
  350. 42.1 NbAiLab/borealis-1b-instruct-preview
  351. 41.9 syvai/claes-v1
  352. 41.9 Gemma 3 1B
  353. 41.9 NbAiLab/borealis-open-1b
  354. 41.6 NbAiLab/notram-bert-norwegian-cased-080321
  355. 41.5 ibm-granite/granite-34b-code-instruct
  356. 41.4 Qwen/Qwen2.5-1.5B-Instruct
  357. 41.3 arcee-ai/AFM-4.5B-Base
  358. 41.3 ibm-granite/granite-4.0-h-tiny
  359. 41.0 Qwen3.5-0.8B
  360. 40.8 jhu-clsp/mmBERT-base
  361. 40.8 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f
  362. 40.7 Qwen/Qwen2.5-Math-7B
  363. 40.7 AI-Sweden-Models/bert-large-nordic-pile-1M-steps
  364. 40.6 CohereForAI/aya-23-8B
  365. 40.5 KennethEnevoldsen/dfm-sentence-encoder-small
  366. 40.4 ibm-granite/granite-8b-code-instruct-4k
  367. 40.4 OpenLLM-France/Luciole-8B-Instruct-1.1
  368. 40.4 Phi-4-reasoning
  369. 40.0 NbAiLab/borealis-1b
  370. 39.7 common-pile/comma-v0.1-2t
  371. 39.6 HuggingFaceTB/SmolLM3-3B-Base
  372. 39.5 swiss-ai/Apertus-v1.1-1.5B
  373. 39.5 FacebookAI/xlm-roberta-large
  374. 39.4 speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551
  375. 39.3 Gemma 2 9B
  376. 39.2 Apertus 8B Instruct
  377. 39.2 utter-project/EuroVLM-1.7B-Preview
  378. 39.1 ibm-granite/granite-34b-code-base
  379. 39.0 meta-llama/Llama-2-7b-chat-hf
  380. 38.9 Gemma 2B
  381. 38.8 infly/OpenCoder-8B-Instruct
  382. 38.6 NbAiLab/nb-bert-large
  383. 38.5 allenai/Olmo-3-7B-Instruct
  384. 38.5 danish-foundation-models/gemma-3-1b-cpt-dynaword-full-v1
  385. 38.5 BSC-LT/salamandra-7b-instruct
  386. 38.4 speakleash/Bielik-1.5B-v3.0-Instruct
  387. 38.4 ltg/norbert3-small
  388. 38.3 ibm-granite/granite-8b-code-base-4k
  389. 38.2 ibm-granite/granite-4.0-h-micro
  390. 38.2 swiss-ai/Apertus-v1.1-1.5B-Instruct
  391. 37.9 allenai/Olmo-3-7B-Instruct-DPO
  392. 37.9 ltg/norbert2
  393. 37.2 danish-foundation-models/munin-7b-alpha
  394. 37.2 KBLab/megatron-bert-base-swedish-cased-600k
  395. 37.1 ibm-granite/granite-20b-code-base
  396. 36.9 norallm/normistral-7b-scratch
  397. 36.8 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660
  398. 36.8 DDSC/roberta-base-danish
  399. 36.7 Gemma 3 1B PT
  400. 36.6 google/gemma-2-2b
  401. 36.5 danish-foundation-models/munin-gemma4-e4b
  402. 36.5 OmniJev/OneJev1.6-0.8B
  403. 36.2 ibm-granite/granite-3.0-3b-a800m-instruct
  404. 36.1 Hunyuan-4B-Instruct
  405. 36.0 common-pile/comma-v0.1-1t
  406. 35.7 syvai/claes-preview
  407. 35.5 LiquidAI/LFM2-2.6B
  408. 35.5 google/rembert
  409. 35.5 KBLab/bert-base-swedish-cased-new
  410. 35.4 Maltehb/danish-bert-botxo
  411. 35.4 danish-foundation-models/munin-7b-core-pt-3
  412. 35.0 HPLT/gpt-13b-nordic-prerelease
  413. 35.0 ibm-granite/granite-3b-code-base-2k
  414. 34.9 HuggingFaceTB/SmolLM3-3B
  415. 34.9 KBLab/megatron-bert-large-swedish-cased-165k
  416. 34.5 allenai/Olmo-3-7B-Instruct-SFT
  417. 34.3 yulan-team/YuLan-Mini-Instruct
  418. 34.3 allenai/OLMo-2-1124-7B
  419. 34.2 danish-foundation-models/munin-7b-core-it
  420. 33.7 tencent/Hunyuan-7B-Instruct
  421. 33.2 fastino/GLiNER2.5-Decide
  422. 33.1 CohereLabs/aya-23-8B
  423. 32.9 microsoft/mdeberta-v3-base
  424. 32.8 ltg/norbert3-xs
  425. 32.7 vesteinn/ScandiBERT
  426. 32.5 mideind/IceBERT-xlmr-ic3
  427. 32.2 MaLA-LM/emma-500-llama2-7b
  428. 31.9 KBLab/megatron-bert-base-swedish-cased-125k
  429. 31.7 TildeAI/TildeOpen-30b#slow-tokenizer
  430. 31.6 HuggingFaceTB/SmolLM3-3B#no-thinking
  431. 31.6 KBLab/bert-base-swedish-cased
  432. 31.5 jhu-clsp/mmBERT-small
  433. 31.5 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6
  434. 31.0 google-bert/bert-base-multilingual-cased
  435. 30.8 danish-foundation-models/munin-7b-core-pt-2
  436. 30.8 Falcon3-3B-Instruct
  437. 30.6 utter-project/EuroLLM-1.7B-Instruct
  438. 30.1 danish-foundation-models/encoder-small-v1
  439. 29.6 google-bert/bert-base-multilingual-uncased
  440. 29.6 TildeAI/TildeOpen-30b
  441. 29.4 allenai/Olmo-3-1125-32B
  442. 29.3 KBLab/roberta-base-swedish-cased
  443. 28.8 nvidia/NVIDIA-Nemotron-Nano-9B-v2#no-thinking
  444. 28.7 norallm/norbloom-7b-scratch
  445. 28.5 tiiuae/Falcon3-3B-Base
  446. 28.2 TurkuNLP/finnish-modernbert-tiny
  447. 28.1 PleIAs/Pleias-1.2b-Preview
  448. 28.0 Llama 3.2 1B
  449. 27.8 Maltehb/aelaectra-danish-electra-small-uncased
  450. 27.8 distilbert/distilbert-base-multilingual-cased
  451. 27.7 Geotrend/distilbert-base-25lang-cased
  452. 27.6 google/gemma-7b-it
  453. 27.4 Qwen3-0.6B
  454. 27.4 KBLab/sentence-bert-swedish-cased
  455. 27.4 meta-llama/Llama-2-13b-hf
  456. 27.3 NorwAI/NorwAI-Mixtral-8x7B-one-third-epoch
  457. 26.9 BSC-LT/ALIA-40b
  458. 26.9 ibm-granite/granite-20b-code-instruct
  459. 26.6 allenai/OLMo-2-0325-32B
  460. 26.4 TurkuNLP/finnish-modernbert-large
  461. 25.9 TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49
  462. 25.8 LFM2.5-1.2B-Thinking
  463. 25.7 convaiinnovations/laya-multilingual
  464. 25.3 TurkuNLP/finnish-modernbert-large-seq-len-1024-117300-annealed
  465. 24.3 Qwen3.5-9B-Base
  466. 24.1 swiss-ai/Apertus-v1.1-0.5B-Instruct
  467. 24.1 swiss-ai/Apertus-v1.1-0.5B
  468. 23.9 Maltehb/aelaectra-danish-electra-small-cased
  469. 23.8 TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894
  470. 23.5 tencent/Hunyuan-1.8B-Instruct
  471. 23.4 EuropeanParliament/EUBERT
  472. 23.2 LFM2.5-1.2B-Instruct
  473. 23.0 neurocode/IsRoBERTa
  474. 22.6 KBLab/bert-base-swedish-cased-alpha
  475. 22.5 OpenLLM-France/Luciole-1B-Instruct-1.1
  476. 22.5 OpenLLM-France/Luciole-23B-Base
  477. 22.5 EuroBERT/EuroBERT-610m
  478. 22.3 EuroBERT/EuroBERT-210m
  479. 22.3 google/gemma-2b-it
  480. 22.2 TurkuNLP/finnish-modernbert-base
  481. 22.1 timpal0l/Mistral-7B-v0.1-flashback-v2
  482. 21.2 vesteinn/XLMR-ENIS
  483. 21.1 arcee-ai/Trinity-Nano-Preview
  484. 20.7 marksverdhei/modern-norbert3-large
  485. 20.6 answerdotai/ModernBERT-base
  486. 20.5 mideind/IceBERT-large
  487. 20.3 danish-foundation-models/munin-7b-core-pt
  488. 20.1 jaredpalmer/kev-0.5b
  489. 20.1 mideind/IceBERT-igc
  490. 20.0 DeepSeek-R1-Distill-Qwen-7B
  491. 19.8 Qwen/Qwen3-0.6B#no-thinking
  492. 19.4 Phi-2
  493. 19.4 EuroBERT/EuroBERT-2.1B
  494. 19.2 Phi-4-mini-reasoning
  495. 19.1 answerdotai/ModernBERT-large
  496. 19.0 arcee-ai/Trinity-Mini-Base
  497. 18.8 speakleash/Bielik-11B-v3-Base-20250730
  498. 18.6 danish-foundation-models/munin-7b-open-pt
  499. 18.2 mideind/IceBERT
  500. 17.5 mideind/IceBERT-ic3
  501. 17.4 DeepSeek-R1-Distill-Qwen-1.5B
  502. 17.4 AI-Sweden-Models/gpt-sw3-126m-instruct
  503. 17.2 LiquidAI/LFM2-350M
  504. 17.2 facebook/MobileLLM-R1-950M-base
  505. 17.1 allenai/OLMo-2-1124-13B
  506. 16.9 AI-Sweden-Models/gpt-sw3-1.3b-instruct
  507. 16.9 LFM2-1.2B
  508. 16.6 PleIAs/Pleias-3b-Preview
  509. 16.5 utter-project/EuroMoE-2.6B-A0.6B-Preview
  510. 16.2 Gemma 3 270M
  511. 16.1 utter-project/EuroLLM-22B-2512
  512. 16.0 NorwAI/NorwAI-Mixtral-8x7B
  513. 16.0 fresh-electra-small
  514. 16.0 fresh-xlm-roberta-base
  515. 15.8 LiquidAI/LFM2.5-1.2B-Base
  516. 15.8 DDSC/roberta-base-scandinavian
  517. 15.7 OpenLLM-France/Luciole-8B-Base
  518. 15.7 SmolLM2 1.7B Instruct
  519. 15.6 DiscoResearch/DiscoLM-70b
  520. 15.4 infly/OpenCoder-8B-Base
  521. 15.3 marksverdhei/modern-norbert3-xs
  522. 15.2 ibm-granite/granite-3b-code-instruct-2k
  523. 15.0 marksverdhei/modern-norbert3-base
  524. 14.8 openGPT-X/Teuken-7B-instruct-commercial-v0.4
  525. 14.3 NbAiLab/roberta_jan_128_ncc
  526. 14.0 convaiinnovations/laya
  527. 13.1 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview
  528. 13.0 RJuro/kanelsnegl-v0.1
  529. 12.9 infly/OpenCoder-1.5B-Instruct
  530. 12.6 LiquidAI/LFM2-700M
  531. 12.4 HPLT/hplt2c_swe_checkpoints
  532. 10.5 PleIAs/Pleias-350m-Preview
  533. 10.5 arcee-ai/Trinity-Nano-Base
  534. 9.75 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base
  535. 9.61 marksverdhei/modern-norbert3-small
  536. 8.98 ibm-granite/granite-3.0-3b-a800m-base
  537. 8.53 SmolLM2-1.7B
  538. 8.14 HuggingFaceTB/SmolLM2-135M-Instruct
  539. 7.71 HPLT/hplt2c_isl_checkpoints
  540. 7.51 mideind/IceBERT-mC4-is
  541. 6.93 Qwen3.5-4B-Base
  542. 6.80 PleIAs/Monad
  543. 6.70 LFM2.5-8B-A1B
  544. 6.42 Phi-4-Reasoning-plus
  545. 6.38 meta-llama/Llama-3.2-1B-Instruct
  546. 6.16 Qwen3.5-2B-Base
  547. 5.74 HPLT/hplt2c_nob_checkpoints
  548. 5.49 NbAiLab/borealis-270m
  549. 5.14 BSC-LT/salamandra-7b
  550. 4.99 SmolLM2-360M
  551. 4.84 NbAiLab/borealis-open-270m
  552. 4.45 Qwen3.5-0.8B-Base
  553. 3.67 openai-community/gpt2
  554. 3.64 AI-Sweden-Models/gpt-sw3-356m-instruct
  555. 3.55 SmolLM2-135M
  556. 3.44 AI-Sweden-Models/gpt-sw3-6.7b-v2-instruct
  557. 3.29 baidu/ERNIE-4.5-0.3B-PT
  558. 2.78 infly/OpenCoder-1.5B-Base
  559. 2.69 AI-Sweden-Models/gpt-sw3-40b
  560. 2.52 openbmb/MiniCPM5-1B
  561. 2.29 openai-community/gpt2-medium
  562. 2.13 NbAiLab/borealis-270m-instruct-preview
  563. 1.77 kyutai/helium-1-2b
  564. 1.74 PleIAs/Baguettotron@refs/pr/6#no-thinking
  565. 1.62 Muse-research/Muse-3B
  566. 1.44 openai-community/gpt2-large
  567. 1.16 OpenLLM-France/Luciole-1B-Base
  568. 0.96 HPLT/hplt2c_dan_checkpoints
  569. 0.77 HuggingFaceTB/SmolLM2-360M-Instruct
  570. 0.50 AI-Sweden-Models/gpt-sw3-6.7b-v2-translator
  571. 0.27 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct
  572. 0.25 facebook/MobileLLM-R1-360M
  573. 0.15 ltg/norbert
  574. 0.12 facebook/MobileLLM-R1-140M
  575. 0.0 AI-Sweden-Models/gpt-sw3-1.3b
  576. 0.0 AI-Sweden-Models/gpt-sw3-20b
  577. 0.0 AI-Sweden-Models/gpt-sw3-356m
  578. 0.0 AI-Sweden-Models/gpt-sw3-6.7b
  579. 0.0 AI-Sweden-Models/gpt-sw3-6.7b-v2
  580. 0.0 facebook/MobileLLM-R1-140M-base
  581. 0.0 facebook/MobileLLM-R1-950M
  582. 0.0 NorwAI/NorwAI-Magistral-24B-reasoning
  583. 0.0 openGPT-X/Teuken-7B-base-v0.6
  584. 0.0 Qwen/Qwen2.5-Math-1.5B
  585. 0.0 utter-project/EuroLLM-1.7B
  586. -0.02 PleIAs/Baguettotron@refs/pr/6#thinking
  587. -0.35 facebook/MobileLLM-R1-360M-base
  588. -0.37 AI-Sweden-Models/gpt-sw3-126m
  589. -0.47 danish-foundation-models/munin-7b-open-it
  590. -0.72 HPLT/hplt2c_nno_checkpoints
  591. -0.89 inclusionAI/Ling-mini-base-2.0
  592. -0.98 Shekswess/trlm-135m
  593. -1.12 PleIAs/Baguettotron@refs/pr/6
  594. -1.15 inclusionAI/Ling-mini-2.0
  595. -1.60 norallm/normistral-11b-translate
  596. -2.03 AIDC-AI/Marco-Mini-Global-Base
  1. 1 GPT-6.2 Astra zero-shot, val 67%
  2. 2 GLM-5.3-Flash val 66.3%
  3. 3 GPT-6 Astra zero-shot, val 65.3%
  4. 4 mistralai/Mistral-Large-3-675B-Instruct-2512 val 64.5%
  5. 5 claude-sonnet-4-5-20250929#thinking zero-shot, val 64.3%
  6. 6 Gemini 3.7 Flash zero-shot, val 64.2%
  7. 7 danish-foundation-models/DFM-Mimir-v1.5 val 64%
  8. 8 Gemini 2.5 Flash Sep 2025, zero-shot, val 63.8%
  9. 9 speakleash/Bielik-11B-v3.1-Instruct val 63.6%
  10. 10 Qwen/Qwen3.5-27B-FP8 val 63.5%
  11. 11 Gemma 4 31B IT val 63.5%
  12. 12 NbAiLab/borealis2-31b-preview val 63.3%
  13. 13 claude-3-7-sonnet-20250219#no-thinking zero-shot, val 63.2%
  14. 14 Qwen/Qwen3.6-27B-FP8 val 63.2%
  15. 15 syvai/danskgpt-v4-31b zero-shot, val 63.2%
  16. 16 deepseek/deepseek-flash#no-thinking val 63.2%
  17. 17 GPT-6.2 Sol zero-shot, val 62.8%
  18. 18 speakleash/Bielik-11B-v3.0-Instruct val 62.6%
  19. 19 claude-sonnet-4-5-20250929#no-thinking zero-shot, val 62.6%
  20. 20 GPT-6.1 Sol zero-shot, val 62.6%
  21. 21 Mistral Medium 3.5 val 62.2%
  22. 22 o3-mini zero-shot, val 62.2%
  23. 23 Gemini 3.5 Flash zero-shot, val 61.9%
  24. 24 gemini/gemini-3-flash-preview#thinking zero-shot, val 61.7%
  25. 25 zai-org/GLM-4.7-FP8 val 61.6%
  26. 26 danish-foundation-models/DFM-Mimir-v1.8 val 61.5%
  27. 27 Qwen3.5 397B-A17B 60.9%
  28. 28 meta-llama/Llama-3.1-405B-Instruct-FP8 60.8%
  29. 29 Claude Opus 4.8 zero-shot, val 60.6%
  30. 30 deepseek/deepseek-flash#low zero-shot, val 60.6%
  31. 31 Qwen 3.8 27B val 60.6%
  32. 32 Qwen/Qwen3.5-35B-A3B-FP8 val 60.5%
  33. 33 GPT-6 Luna zero-shot, val 60.2%
  34. 34 Gemini 3.6 Flash zero-shot, val 60.2%
  35. 35 Gemini 3 Pro zero-shot, val 60.2%
  36. 36 Olmo 3.1 32B Think val 60%
  37. 37 Gemma 4 12B val 60%
  38. 38 NbAiLab/borealis-open-27b val 59.8%
  39. 39 danish-foundation-models/DFM-Mimir val 59.6%
  40. 40 NbAiLab/borealis-27b-instruct-preview val 59.5%
  41. 41 GPT-5.6 Sol zero-shot, val 59.3%
  42. 42 Qwen/Qwen3-235B-A22B-Thinking-2507-FP8 59.3%
  43. 43 LumiOpen/Llama-Poro-2-70B-SFT 59.1%
  44. 44 mistralai/Mistral-Small-3.1-24B-Instruct-2503 zero-shot, val 59.1%
  45. 45 ThatsGroes/gemma-2-27b-it-FP8-Dynamic 58.9%
  46. 46 GPT-5 zero-shot, val 58.8%
  47. 47 deepcogito/cogito-v2-preview-llama-70B 58.8%
  48. 48 gemini/gemini-3-flash-preview#no-thinking zero-shot, val 58.7%
  49. 49 Reka Flash 3.1 58.7%
  50. 50 NbAiLab/borealis-27b val 58.6%
  51. 51 Cogito v1 Preview 70B 58.6%
  52. 52 LumiOpen/Llama-Poro-2-70B-Instruct 58.5%
  53. 53 GPT-5.2 zero-shot, val 58.5%
  54. 54 glm-4.7-flash val 58.5%
  55. 55 Llama 3.1-70B 58.4%
  56. 56 claude-3-7-sonnet-20250219#thinking zero-shot, val 58.3%
  57. 57 mistralai/Mistral-Small-4-119B-2603 val 58.3%
  58. 58 Qwen3.5-9B val 58.2%
  59. 59 GadflyII/GLM-4.7-Flash-NVFP4 val 58.2%
  60. 60 openai/gpt-5.4-mini-2026-03-17#medium zero-shot, val 58.1%
  61. 61 Gemini 3.1 Flash-Lite zero-shot, val 58.1%
  62. 62 o3 zero-shot, val 58.1%
  63. 63 Llama 4 Scout 58.1%
  64. 64 Reka Flash 3 58%
  65. 65 EXAONE 4.0 32B val 58%
  66. 66 syvai/danskgpt-v4-12b zero-shot, val 57.9%
  67. 67 Claude Sonnet 4.6 zero-shot, val 57.7%
  68. 68 google/gemma-2-27b-it 57.6%
  69. 69 o4-mini zero-shot, val 57.5%
  70. 70 Gemma 3 27B 57.4%
  71. 71 mistralai/Mistral-Small-24B-Instruct-2501 57.3%
  72. 72 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 val 57.1%
  73. 73 GPT-5.5 zero-shot, val 57%
  74. 74 speakleash/Bielik-11B-v2.3-Instruct val 57%
  75. 75 typesafe/jev-1.13.0 zero-shot 57%
  76. 76 Gemma 4 26B A4B val 56.9%
  77. 77 GPT-4.1 zero-shot, val 56.9%
  78. 78 Nemotron-Cascade 2 30B-A3B val 56.9%
  79. 79 Llama 3.3 70B Instruct zero-shot, val 56.8%
  80. 80 Unbabel/Tower-Plus-72B val 56.8%
  81. 81 mistralai/Mistral-Small-3.2-24B-Instruct-2506 56.8%
  82. 82 gpt-5-nano-2025-08-07#high zero-shot, val 56.7%
  83. 83 mistralai/Mistral-Small-3.1-24B-Base-2503 56.7%
  84. 84 Gemma 3 12B 56.7%
  85. 85 QwQ-32B 56.6%
  86. 86 ollama_chat/gemma3:27b-it-fp16 56.4%
  87. 87 Apertus 1.5 70B val 56.4%
  88. 88 NbAiLab/borealis-12b-instruct-preview val 56.3%
  89. 89 OLMo 2 32B Instruct 56.2%
  90. 90 danish-foundation-models/munin-qwen3.5-9B 56.2%
  91. 91 GLM-4.5-Air thinking 56.2%
  92. 92 gpt-5-2025-08-07#high zero-shot, val 56.2%
  93. 93 ollama_chat/gemma3:12b-it-q4_K_M val 56.2%
  94. 94 deepcogito/cogito-v1-preview-qwen-32B 56.1%
  95. 95 Qwen3.5-35B-A3B-Base val 56.1%
  96. 96 NbAiLab/borealis2-26b-a4b-preview val 56.1%
  97. 97 ollama_chat/gemma3:27b-it-q4_K_M val 56%
  98. 98 Aya Expanse 8B val 56%
  99. 99 GPT-5 nano zero-shot, val 56%
  100. 100 mistralai/Magistral-Small-2509 56%
  101. 101 Qwen/Qwen3-235B-A22B-Instruct-2507-FP8 55.9%
  102. 102 Claude Haiku 4.5 zero-shot, val 55.9%
  103. 103 Grok 4.20 thinking, zero-shot, val 55.9%
  104. 104 openai/gpt-5.4-2026-03-05#none zero-shot, val 55.7%
  105. 105 NovaSky-AI/Sky-T1-32B-Preview 55.7%
  106. 106 Qwen/Qwen3.5-9B#no-thinking 55.7%
  107. 107 Gemini 2.5 Pro Jun 2025, zero-shot, val 55.7%
  108. 108 zai-org/GLM-4.5-Air#no-thinking 55.6%
  109. 109 gpt-oss-120b val 55.6%
  110. 110 Qwen3-14B 55.5%
  111. 111 NbAiLab/borealis2-e4b-preview val 55.5%
  112. 112 Ministral 3 3B Instruct zero-shot, val 55.5%
  113. 113 GPT-5.6 Terra zero-shot, val 55.5%
  114. 114 Apertus 70B 55.4%
  115. 115 Gemma 3n E4B 55.4%
  116. 116 Qwen3-235B-A22B 55.4%
  117. 117 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8 val 55.3%
  118. 118 Qwen/Qwen3-14B#no-thinking 55.2%
  119. 119 Aleph-Alpha/Kolibri-1 val 55.2%
  120. 120 meta-llama/Llama-3.1-8B-Instruct 55.2%
  121. 121 Gemma 3n E2B 55.2%
  122. 122 GPT-6 Sol zero-shot, val 55.2%
  123. 123 Gemini 3.5 Flash-Lite zero-shot, val 55.1%
  124. 124 Qwen3-Next-80B-A3B thinking 55.1%
  125. 125 openai/gpt-5.4-mini-2026-03-17#high zero-shot, val 55.1%
  126. 126 NbAiLab/borealis-open-12b val 55%
  127. 127 Grok-2 Dec 2024, zero-shot, val 55%
  128. 128 Phi-4 55%
  129. 129 deepcogito/cogito-v1-preview-qwen-14B 54.9%
  130. 130 Grok 3 zero-shot, val 54.8%
  131. 131 speakleash/Bielik-Minitron-7B-v3.0-Instruct val 54.7%
  132. 132 Qwen/Qwen3-32B#no-thinking 54.6%
  133. 133 Voxtral Small 54.6%
  134. 134 Qwen3-8B 54.5%
  135. 135 gpt-oss-20b high 54.5%
  136. 136 Grok-3 mini zero-shot, val 54.5%
  137. 137 NousResearch/DeepHermes-3-Llama-3-8B-Preview 54.4%
  138. 138 utter-project/EuroLLM-9B-Instruct 54.4%
  139. 139 Gemma 3 27B PT 54.3%
  140. 140 deepcogito/cogito-v1-preview-llama-8B 54.2%
  141. 141 NbAiLab/borealis-12b val 54.2%
  142. 142 timpal0l/BeagleCatMunin val 54.1%
  143. 143 Grok 4.1 Fast zero-shot, val 54.1%
  144. 144 Aya Expanse 32B 54.1%
  145. 145 alexandrainst/Meta-Llama-3.1-8B-laerebogen-and-dialog-k2-da val 54%
  146. 146 GLM-4.1V-9B-Thinking 54%
  147. 147 openai/gpt-5.4-mini-2026-03-17#low zero-shot, val 53.9%
  148. 148 ollama_chat/gemma3:12b-it-fp16 val 53.6%
  149. 149 gpt-5-2025-08-07#minimal zero-shot, val 53.6%
  150. 150 alexandrainst/Meta-Llama-3.1-8B-laerebogen val 53.4%
  151. 151 gemini/gemini-2.5-flash#no-thinking zero-shot, val 53.4%
  152. 152 Qwen3-Next-80B-A3B-Instruct 53.3%
  153. 153 ollama_chat/phi4:14b-q4_K_M val 53.3%
  154. 154 allenai/Olmo-3.1-32B-Instruct-DPO 53.2%
  155. 155 KennethEnevoldsen/dfm-sentence-encoder-large 53.2%
  156. 156 gemini/gemini-2.5-flash-lite-preview-06-17#thinking zero-shot, val 53.2%
  157. 157 Gemini 2.5 Flash-Lite Sep 2025, zero-shot, val 53.2%
  158. 158 google/gemma-2-9b-it 53.2%
  159. 159 utter-project/EuroLLM-22B-Instruct-2512 53.1%
  160. 160 gpt-5-mini-2025-08-07#high zero-shot, val 53%
  161. 161 Gemma 3 12B PT 53%
  162. 162 utter-project/EuroLLM-9B 53%
  163. 163 allenai/Llama-3.1-Tulu-3-8B-DPO 52.9%
  164. 164 intfloat/multilingual-e5-large 52.9%
  165. 165 mistralai/Ministral-3-14B-Base-2512 val 52.9%
  166. 166 allenai/Olmo-3.1-32B-Instruct-SFT 52.8%
  167. 167 Ministral 3 14B val 52.8%
  168. 168 danish-foundation-models/munin-ministral3-8B 52.8%
  169. 169 DeepSeek-R1-Distill-Llama-70B 52.8%
  170. 170 Gemma 4 E4B val 52.8%
  171. 171 GPT-5 mini zero-shot, val 52.7%
  172. 172 ordbogen/odin-medium zero-shot, val 52.7%
  173. 173 Qwen/Qwen3-30B-A3B#no-thinking 52.7%
  174. 174 timpal0l/dolphin-2.9-llama3-8b-flashback val 52.6%
  175. 175 Llama-3.1-Tulu-3-8B 52.6%
  176. 176 GPT-5.6 Luna zero-shot, val 52.5%
  177. 177 merge-crew/da-sv-ties val 52.5%
  178. 178 utter-project/EuroLLM-22B-Instruct-Preview 52.5%
  179. 179 Olmo 3.1 32B Instruct 52.3%
  180. 180 alexandrainst/Meta-Llama-3.1-8B-laerebogen-two-epochs val 52.3%
  181. 181 01-ai/Yi-1.5-9B 52.3%
  182. 182 birgermoell/BeagleCatMunin-Flashback-Bellman val 52.2%
  183. 183 mistralai/Ministral-3-14B-Instruct-2512 val 52.2%
  184. 184 Command R7B 52.1%
  185. 185 Qwen3-4B-Thinking-2507 52.1%
  186. 186 ordbogen/odin-large zero-shot, val 52.1%
  187. 187 RJuro/munin-neuralbeagle-7b val 52%
  188. 188 meta-llama/Llama-4-Scout-17B-16E 52%
  189. 189 speakleash/Bielik-11B-v2.6-Instruct 52%
  190. 190 bineric/lynx-instruct-30b val 51.9%
  191. 191 Llama 3.1-405B 51.9%
  192. 192 merge-crew/da-sv-dare-ties-density-0.9 val 51.9%
  193. 193 openGPT-X/Teuken-7B-instruct-v0.6 val 51.8%
  194. 194 merge-crew/da-sv-task-arithmetic val 51.8%
  195. 195 Granite 4.0 H Small 51.8%
  196. 196 birgermoell/Flashback-Bellman val 51.6%
  197. 197 birgermoell/Rapid-Cycling val 51.6%
  198. 198 TrustLLMeu/baseline-7-8b_2-3t-tokens_llama 51.6%
  199. 199 Qwen/Qwen3-8B#no-thinking 51.6%
  200. 200 openai/gpt-5.4-mini-2026-03-17#none zero-shot, val 51.6%
  201. 201 merge-crew/da-sv-slerp val 51.5%
  202. 202 gemini/gemini-2.5-flash-lite-preview-06-17#no-thinking zero-shot, val 51.5%
  203. 203 gemini/gemini-2.5-flash-lite#no-thinking zero-shot, val 51.5%
  204. 204 CohereForAI/aya-23-35B 51.3%
  205. 205 openai/gpt-5.4-nano-2026-03-17#medium zero-shot, val 51.3%
  206. 206 RJuro/munin-neuralbeagle-SkoleGPTOpenOrca-7b val 51.1%
  207. 207 LumiOpen/Viking-33B 51.1%
  208. 208 bullerwins/Magistral-Small-2506-fp8 51.1%
  209. 209 Trinity Mini 51.1%
  210. 210 Qwen1.5-14B-Chat 51%
  211. 211 danish-foundation-models/Meta-Llama-3.1-8B-laerebogen 50.9%
  212. 212 tiiuae/falcon-11B 50.9%
  213. 213 Granite-3.0-8B-Instruct 50.8%
  214. 214 Qwen3.5-4B val 50.8%
  215. 215 NbAiLab/nb-roberta-base-scandi 50.7%
  216. 216 openai/gpt-5.4-nano-2026-03-17#low zero-shot, val 50.7%
  217. 217 danish-foundation-models/DFM-Mimir-v1.8-Mini zero-shot, val 50.7%
  218. 218 Unbabel/Tower-Plus-9B val 50.6%
  219. 219 allenai/Llama-3.1-Tulu-3.1-8B 50.5%
  220. 220 alexandrainst/Meta-Llama-3.1-8B-laerebogen-and-filtered-dialog-k2-da val 50.5%
  221. 221 meta-llama/Llama-2-70b-hf 50.5%
  222. 222 ollama_chat/hf.co/google/gemma-3-4b-it-qat-q4_0-gguf 50.5%
  223. 223 ollama_chat/gemma3:4b-it-q4_K_M 50.3%
  224. 224 openai/gpt-5.4-nano-2026-03-17#none zero-shot, val 50.3%
  225. 225 senseable/WestLake-7B-v2 50.3%
  226. 226 danish-foundation-models/encoder-large-v1 50.3%
  227. 227 merge-crew/da-sv-dare-ties-density-0.6 val 50.3%
  228. 228 152334H/miqu-1-70b-sf 50.2%
  229. 229 mistralai/Ministral-3-8B-Reasoning-2512 50.1%
  230. 230 deepcogito/cogito-v1-preview-llama-3B 50%
  231. 231 KennethEnevoldsen/dfm-sentence-encoder-medium 50%
  232. 232 intfloat/multilingual-e5-base 49.8%
  233. 233 meta-llama/Llama-2-70b-chat-hf 49.8%
  234. 234 NbAiLab/nb-roberta-base-scandi-1e4 49.7%
  235. 235 AI-Sweden-Models/Llama-3-8B-instruct val 49.7%
  236. 236 NbAiLab/nb-roberta-base-ncc-plus-scandi-1e4 49.7%
  237. 237 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-Base-BF16 val 49.7%
  238. 238 AI-Sweden-Models/roberta-large-1160k 49.6%
  239. 239 AI-Sweden-Models/tyr val 49.6%
  240. 240 openai/gpt-5.4-nano-2026-03-17#high zero-shot, val 49.6%
  241. 241 Gemma 4 E2B val 49.5%
  242. 242 NbAiLab/borealis-4b val 49.5%
  243. 243 utter-project/EuroLLM-22B-Preview 49.5%
  244. 244 DeepSeek-R1-Distill-Qwen-32B 49.5%
  245. 245 NbAiLab/nb-roberta-base-ncc-plus-scandi-2e4 49.5%
  246. 246 Ministral 3 8B Instruct val 49.4%
  247. 247 OpenLLM-France/Luciole-23B-Instruct-1.1 val 49.3%
  248. 248 Qwen/Qwen3-4B#no-thinking 49.3%
  249. 249 Qwen3-4B 49.3%
  250. 250 utter-project/EuroLLM-9B-Instruct-2512 val 49.3%
  251. 251 LumiOpen/Viking-13B 49.3%
  252. 252 Unbabel/Tower-Plus-2B val 49.2%
  253. 253 upstage/SOLAR-10.7B-v1.0 49.1%
  254. 254 gpt-5-nano-2025-08-07#minimal zero-shot, val 49.1%
  255. 255 KennethEnevoldsen/munin_mistral-7b val 49.1%
  256. 256 Falcon3-10B-Instruct 49.1%
  257. 257 deepseek-r1-0528-qwen3-8b 49%
  258. 258 Gemma 2 27B 49%
  259. 259 gpt-5-mini-2025-08-07#minimal zero-shot, val 49%
  260. 260 NbAiLab/borealis2-e2b-preview val 48.9%
  261. 261 timpal0l/BeagleCatMunin2 val 48.8%
  262. 262 Ministral 3 3B Reasoning 48.7%
  263. 263 amalia-llm/AMALIA-9B-0626-DPO zero-shot, val 48.7%
  264. 264 danish-foundation-models/munin-apertus-8b 48.7%
  265. 265 synquid/gemma-3-4b-pt-laerebogen-with-follow-ups-lora 48.6%
  266. 266 Tiny Aya Water val 48.6%
  267. 267 allenai/Olmo-3-7B-Think 48.5%
  268. 268 meta-llama/Llama-2-13b-chat-hf 48.4%
  269. 269 NbAiLab/borealis-4b-instruct-preview 48.2%
  270. 270 NbAiLab/borealis-open-4b val 48.2%
  271. 271 Granite-3.0-8B-Base 48.1%
  272. 272 CohereLabs/tiny-aya-base val 48.1%
  273. 273 vesteinn/DanskBERT 48.1%
  274. 274 amalia-llm/AMALIA-9B-0626-SFT val 48%
  275. 275 Llama 3.1-8B 48%
  276. 276 mistralai/Ministral-3-3B-Base-2512 48%
  277. 277 nvidia-nemotron-nano-9b-v2 48%
  278. 278 granite-4.0-micro 48%
  279. 279 Tiny Aya Global val 47.9%
  280. 280 Falcon3-7B-Instruct 47.8%
  281. 281 swiss-ai/Apertus-v1.1-4B val 47.7%
  282. 282 allenai/OLMo-2-1124-13B-Instruct 47.7%
  283. 283 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16#no-thinking 47.6%
  284. 284 mistralai/Ministral-3-8B-Base-2512 val 47.6%
  285. 285 neph1/bellman-7b-mistral-instruct-v0.2 47.6%
  286. 286 allenai/Llama-3.1-Tulu-3-8B-SFT 47.5%
  287. 287 sentence-transformers/use-cmlm-multilingual 47.5%
  288. 288 tiiuae/Falcon3-10B-Base 47.5%
  289. 289 qwen3-4b-instruct-2507 47.4%
  290. 290 Mistral 7B v0.3 47.4%
  291. 291 mlabonne/NeuralBeagle14-7B val 47.4%
  292. 292 vesteinn/FoBERT 47.4%
  293. 293 Gemma 7B 47.3%
  294. 294 openeurollm/prelude-long val 47.3%
  295. 295 occiglot/occiglot-7b-eu5 47.2%
  296. 296 bineric/lynx-instruct-30b#no-thinking 47.2%
  297. 297 Gemma 2 2B IT 47.1%
  298. 298 domyn/Domyn-Small-v1.0 val 47.1%
  299. 299 DGurgurov/xlm-r_dan-latn 46.9%
  300. 300 norallm/normistral-11b-long 46.9%
  301. 301 bineric/lynx-instruct-30b#thinking 46.8%
  302. 302 openeurollm/oellm-9b-128k-theta32m-v3 val 46.8%
  303. 303 openeurollm/prelude-base val 46.7%
  304. 304 Yi-1.5-34B 46.6%
  305. 305 mistralai/Ministral-8B-Instruct-2410 46.3%
  306. 306 mii-llm/nesso-4B 46.2%
  307. 307 swiss-ai/Apertus-v1.1-4B-Instruct val 46.2%
  308. 308 vesteinn/ScandiBERT-no-faroese 46.2%
  309. 309 nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16 46.1%
  310. 310 DeepSeek-R1-Distill-Qwen-14B 45.9%
  311. 311 Apertus 1.5 8B zero-shot, val 45.8%
  312. 312 Qwen3-1.7B 45.7%
  313. 313 allenai/OLMo-2-1124-7B-Instruct 45.6%
  314. 314 Qwen/Qwen3-1.7B#no-thinking 45.5%
  315. 315 Mistral 7B v0.1 45.2%
  316. 316 alexandrainst/Meta-Llama-3.1-8B-dialog-k2-da zero-shot, val 45.1%
  317. 317 TrustLLMeu/baseline-7-8b_1t-tokens_llama 45.1%
  318. 318 HPLT/NorOLMo-13B val 45%
  319. 319 Mistral-7B-Instruct-v0.2 44.9%
  320. 320 Mistral-7B-Instruct-v0.1 44.7%
  321. 321 LiquidAI/LFM2.5-8B-A1B-Base val 44.6%
  322. 322 tiiuae/Falcon3-7B-Base 44.6%
  323. 323 birgermoell/Munin-NeuralBeagle-NorskGPT val 44.5%
  324. 324 ai21labs/Jamba-v0.1 44.5%
  325. 325 RuterNorway/Llama-2-13b-chat-norwegian 44.5%
  326. 326 norallm/normistral-11b-thinking 44.4%
  327. 327 baidu/ERNIE-4.5-21B-A3B-PT 44.3%
  328. 328 allenai/Olmo-3-1025-7B 44.3%
  329. 329 mistralai/Voxtral-Mini-3B-2507 44.2%
  330. 330 swiss-ai/Apertus-8B-2509 44.1%
  331. 331 setu4993/LaBSE 44.1%
  332. 332 AI-Sweden-Models/ModernBERT-large 44%
  333. 333 ltg/norbert3-large 44%
  334. 334 deepseek-ai/DeepSeek-R1-Distill-Llama-8B 43.9%
  335. 335 nvidia/NVIDIA-Nemotron-3-Nano-4B-FP8 val 43.8%
  336. 336 NbAiLab/nb-bert-ncc-male2female 43.8%
  337. 337 NbAiLab/nb-roberta-base-ncc 43.7%
  338. 338 Qwen1.5-14B 43.4%
  339. 339 norallm/normistral-11b-warm 43.1%
  340. 340 deahmed/Qwen3.5-2B-da-task val 43%
  341. 341 speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c zero-shot 43%
  342. 342 NbAiLab/nb-bert-base 42.9%
  343. 343 Qwen3.5-2B val 42.7%
  344. 344 mlabonne/AlphaMonarch-7B val 42.6%
  345. 345 speakleash/Bielik-4.5B-v3.0-Instruct 42.6%
  346. 346 occiglot/occiglot-7b-eu5-instruct 42.5%
  347. 347 AFM-4.5B 42.5%
  348. 348 ltg/norbert3-base 42.3%
  349. 349 meta-llama/Llama-3.2-3B-Instruct 42.2%
  350. 350 NbAiLab/borealis-1b-instruct-preview val 42.1%
  351. 351 syvai/claes-v1 zero-shot 41.9%
  352. 352 Gemma 3 1B 41.9%
  353. 353 NbAiLab/borealis-open-1b val 41.9%
  354. 354 NbAiLab/notram-bert-norwegian-cased-080321 41.6%
  355. 355 ibm-granite/granite-34b-code-instruct 41.5%
  356. 356 Qwen/Qwen2.5-1.5B-Instruct 41.4%
  357. 357 arcee-ai/AFM-4.5B-Base 41.3%
  358. 358 ibm-granite/granite-4.0-h-tiny 41.3%
  359. 359 Qwen3.5-0.8B val 41%
  360. 360 jhu-clsp/mmBERT-base 40.8%
  361. 361 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f 40.8%
  362. 362 Qwen/Qwen2.5-Math-7B 40.7%
  363. 363 AI-Sweden-Models/bert-large-nordic-pile-1M-steps 40.7%
  364. 364 CohereForAI/aya-23-8B 40.6%
  365. 365 KennethEnevoldsen/dfm-sentence-encoder-small 40.5%
  366. 366 ibm-granite/granite-8b-code-instruct-4k 40.4%
  367. 367 OpenLLM-France/Luciole-8B-Instruct-1.1 val 40.4%
  368. 368 Phi-4-reasoning 40.4%
  369. 369 NbAiLab/borealis-1b val 40%
  370. 370 common-pile/comma-v0.1-2t 39.7%
  371. 371 HuggingFaceTB/SmolLM3-3B-Base 39.6%
  372. 372 swiss-ai/Apertus-v1.1-1.5B val 39.5%
  373. 373 FacebookAI/xlm-roberta-large 39.5%
  374. 374 speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551 zero-shot 39.4%
  375. 375 Gemma 2 9B 39.3%
  376. 376 Apertus 8B Instruct 39.2%
  377. 377 utter-project/EuroVLM-1.7B-Preview 39.2%
  378. 378 ibm-granite/granite-34b-code-base 39.1%
  379. 379 meta-llama/Llama-2-7b-chat-hf 39%
  380. 380 Gemma 2B 38.9%
  381. 381 infly/OpenCoder-8B-Instruct 38.8%
  382. 382 NbAiLab/nb-bert-large 38.6%
  383. 383 allenai/Olmo-3-7B-Instruct 38.5%
  384. 384 danish-foundation-models/gemma-3-1b-cpt-dynaword-full-v1 38.5%
  385. 385 BSC-LT/salamandra-7b-instruct 38.5%
  386. 386 speakleash/Bielik-1.5B-v3.0-Instruct 38.4%
  387. 387 ltg/norbert3-small 38.4%
  388. 388 ibm-granite/granite-8b-code-base-4k 38.3%
  389. 389 ibm-granite/granite-4.0-h-micro 38.2%
  390. 390 swiss-ai/Apertus-v1.1-1.5B-Instruct val 38.2%
  391. 391 allenai/Olmo-3-7B-Instruct-DPO 37.9%
  392. 392 ltg/norbert2 37.9%
  393. 393 danish-foundation-models/munin-7b-alpha 37.2%
  394. 394 KBLab/megatron-bert-base-swedish-cased-600k 37.2%
  395. 395 ibm-granite/granite-20b-code-base 37.1%
  396. 396 norallm/normistral-7b-scratch val 36.9%
  397. 397 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660 36.8%
  398. 398 DDSC/roberta-base-danish 36.8%
  399. 399 Gemma 3 1B PT 36.7%
  400. 400 google/gemma-2-2b 36.6%
  401. 401 danish-foundation-models/munin-gemma4-e4b 36.5%
  402. 402 OmniJev/OneJev1.6-0.8B zero-shot 36.5%
  403. 403 ibm-granite/granite-3.0-3b-a800m-instruct 36.2%
  404. 404 Hunyuan-4B-Instruct 36.1%
  405. 405 common-pile/comma-v0.1-1t 36%
  406. 406 syvai/claes-preview zero-shot 35.7%
  407. 407 LiquidAI/LFM2-2.6B 35.5%
  408. 408 google/rembert 35.5%
  409. 409 KBLab/bert-base-swedish-cased-new 35.5%
  410. 410 Maltehb/danish-bert-botxo 35.4%
  411. 411 danish-foundation-models/munin-7b-core-pt-3 35.4%
  412. 412 HPLT/gpt-13b-nordic-prerelease 35%
  413. 413 ibm-granite/granite-3b-code-base-2k 35%
  414. 414 HuggingFaceTB/SmolLM3-3B 34.9%
  415. 415 KBLab/megatron-bert-large-swedish-cased-165k 34.9%
  416. 416 allenai/Olmo-3-7B-Instruct-SFT 34.5%
  417. 417 yulan-team/YuLan-Mini-Instruct 34.3%
  418. 418 allenai/OLMo-2-1124-7B 34.3%
  419. 419 danish-foundation-models/munin-7b-core-it 34.2%
  420. 420 tencent/Hunyuan-7B-Instruct 33.7%
  421. 421 fastino/GLiNER2.5-Decide zero-shot 33.2%
  422. 422 CohereLabs/aya-23-8B val 33.1%
  423. 423 microsoft/mdeberta-v3-base 32.9%
  424. 424 ltg/norbert3-xs 32.8%
  425. 425 vesteinn/ScandiBERT 32.7%
  426. 426 mideind/IceBERT-xlmr-ic3 32.5%
  427. 427 MaLA-LM/emma-500-llama2-7b 32.2%
  428. 428 KBLab/megatron-bert-base-swedish-cased-125k 31.9%
  429. 429 TildeAI/TildeOpen-30b#slow-tokenizer 31.7%
  430. 430 HuggingFaceTB/SmolLM3-3B#no-thinking 31.6%
  431. 431 KBLab/bert-base-swedish-cased 31.6%
  432. 432 jhu-clsp/mmBERT-small 31.5%
  433. 433 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6 31.5%
  434. 434 google-bert/bert-base-multilingual-cased val 31%
  435. 435 danish-foundation-models/munin-7b-core-pt-2 30.8%
  436. 436 Falcon3-3B-Instruct 30.8%
  437. 437 utter-project/EuroLLM-1.7B-Instruct val 30.6%
  438. 438 danish-foundation-models/encoder-small-v1 30.1%
  439. 439 google-bert/bert-base-multilingual-uncased val 29.6%
  440. 440 TildeAI/TildeOpen-30b val 29.6%
  441. 441 allenai/Olmo-3-1125-32B 29.4%
  442. 442 KBLab/roberta-base-swedish-cased 29.3%
  443. 443 nvidia/NVIDIA-Nemotron-Nano-9B-v2#no-thinking 28.8%
  444. 444 norallm/norbloom-7b-scratch val 28.7%
  445. 445 tiiuae/Falcon3-3B-Base 28.5%
  446. 446 TurkuNLP/finnish-modernbert-tiny 28.2%
  447. 447 PleIAs/Pleias-1.2b-Preview 28.1%
  448. 448 Llama 3.2 1B 28%
  449. 449 Maltehb/aelaectra-danish-electra-small-uncased 27.8%
  450. 450 distilbert/distilbert-base-multilingual-cased 27.8%
  451. 451 Geotrend/distilbert-base-25lang-cased 27.7%
  452. 452 google/gemma-7b-it 27.6%
  453. 453 Qwen3-0.6B 27.4%
  454. 454 KBLab/sentence-bert-swedish-cased 27.4%
  455. 455 meta-llama/Llama-2-13b-hf 27.4%
  456. 456 NorwAI/NorwAI-Mixtral-8x7B-one-third-epoch val 27.3%
  457. 457 BSC-LT/ALIA-40b 26.9%
  458. 458 ibm-granite/granite-20b-code-instruct 26.9%
  459. 459 allenai/OLMo-2-0325-32B 26.6%
  460. 460 TurkuNLP/finnish-modernbert-large 26.4%
  461. 461 TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49 25.9%
  462. 462 LFM2.5-1.2B-Thinking 25.8%
  463. 463 convaiinnovations/laya-multilingual zero-shot 25.7%
  464. 464 TurkuNLP/finnish-modernbert-large-seq-len-1024-117300-annealed 25.3%
  465. 465 Qwen3.5-9B-Base val 24.3%
  466. 466 swiss-ai/Apertus-v1.1-0.5B-Instruct val 24.1%
  467. 467 swiss-ai/Apertus-v1.1-0.5B val 24.1%
  468. 468 Maltehb/aelaectra-danish-electra-small-cased 23.9%
  469. 469 TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894 23.8%
  470. 470 tencent/Hunyuan-1.8B-Instruct 23.5%
  471. 471 EuropeanParliament/EUBERT 23.4%
  472. 472 LFM2.5-1.2B-Instruct 23.2%
  473. 473 neurocode/IsRoBERTa 23%
  474. 474 KBLab/bert-base-swedish-cased-alpha 22.6%
  475. 475 OpenLLM-France/Luciole-1B-Instruct-1.1 val 22.5%
  476. 476 OpenLLM-France/Luciole-23B-Base val 22.5%
  477. 477 EuroBERT/EuroBERT-610m 22.5%
  478. 478 EuroBERT/EuroBERT-210m 22.3%
  479. 479 google/gemma-2b-it 22.3%
  480. 480 TurkuNLP/finnish-modernbert-base 22.2%
  481. 481 timpal0l/Mistral-7B-v0.1-flashback-v2 22.1%
  482. 482 vesteinn/XLMR-ENIS 21.2%
  483. 483 arcee-ai/Trinity-Nano-Preview 21.1%
  484. 484 marksverdhei/modern-norbert3-large 20.7%
  485. 485 answerdotai/ModernBERT-base 20.6%
  486. 486 mideind/IceBERT-large 20.5%
  487. 487 danish-foundation-models/munin-7b-core-pt 20.3%
  488. 488 jaredpalmer/kev-0.5b zero-shot 20.1%
  489. 489 mideind/IceBERT-igc 20.1%
  490. 490 DeepSeek-R1-Distill-Qwen-7B 20%
  491. 491 Qwen/Qwen3-0.6B#no-thinking 19.8%
  492. 492 Phi-2 19.4%
  493. 493 EuroBERT/EuroBERT-2.1B 19.4%
  494. 494 Phi-4-mini-reasoning 19.2%
  495. 495 answerdotai/ModernBERT-large 19.1%
  496. 496 arcee-ai/Trinity-Mini-Base 19%
  497. 497 speakleash/Bielik-11B-v3-Base-20250730 18.8%
  498. 498 danish-foundation-models/munin-7b-open-pt 18.6%
  499. 499 mideind/IceBERT 18.2%
  500. 500 mideind/IceBERT-ic3 17.5%
  501. 501 DeepSeek-R1-Distill-Qwen-1.5B 17.4%
  502. 502 AI-Sweden-Models/gpt-sw3-126m-instruct 17.4%
  503. 503 LiquidAI/LFM2-350M 17.2%
  504. 504 facebook/MobileLLM-R1-950M-base 17.2%
  505. 505 allenai/OLMo-2-1124-13B 17.1%
  506. 506 AI-Sweden-Models/gpt-sw3-1.3b-instruct val 16.9%
  507. 507 LFM2-1.2B 16.9%
  508. 508 PleIAs/Pleias-3b-Preview 16.6%
  509. 509 utter-project/EuroMoE-2.6B-A0.6B-Preview 16.5%
  510. 510 Gemma 3 270M 16.2%
  511. 511 utter-project/EuroLLM-22B-2512 16.1%
  512. 512 NorwAI/NorwAI-Mixtral-8x7B val 16%
  513. 513 fresh-electra-small 16%
  514. 514 fresh-xlm-roberta-base 16%
  515. 515 LiquidAI/LFM2.5-1.2B-Base 15.8%
  516. 516 DDSC/roberta-base-scandinavian 15.8%
  517. 517 OpenLLM-France/Luciole-8B-Base val 15.7%
  518. 518 SmolLM2 1.7B Instruct 15.7%
  519. 519 DiscoResearch/DiscoLM-70b 15.6%
  520. 520 infly/OpenCoder-8B-Base 15.4%
  521. 521 marksverdhei/modern-norbert3-xs 15.3%
  522. 522 ibm-granite/granite-3b-code-instruct-2k 15.2%
  523. 523 marksverdhei/modern-norbert3-base 15%
  524. 524 openGPT-X/Teuken-7B-instruct-commercial-v0.4 14.8%
  525. 525 NbAiLab/roberta_jan_128_ncc 14.3%
  526. 526 convaiinnovations/laya zero-shot 14%
  527. 527 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview 13.1%
  528. 528 RJuro/kanelsnegl-v0.1 13%
  529. 529 infly/OpenCoder-1.5B-Instruct 12.9%
  530. 530 LiquidAI/LFM2-700M 12.6%
  531. 531 HPLT/hplt2c_swe_checkpoints 12.4%
  532. 532 PleIAs/Pleias-350m-Preview 10.5%
  533. 533 arcee-ai/Trinity-Nano-Base 10.5%
  534. 534 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base 9.75%
  535. 535 marksverdhei/modern-norbert3-small 9.61%
  536. 536 ibm-granite/granite-3.0-3b-a800m-base 8.98%
  537. 537 SmolLM2-1.7B 8.53%
  538. 538 HuggingFaceTB/SmolLM2-135M-Instruct 8.14%
  539. 539 HPLT/hplt2c_isl_checkpoints 7.71%
  540. 540 mideind/IceBERT-mC4-is 7.51%
  541. 541 Qwen3.5-4B-Base val 6.93%
  542. 542 PleIAs/Monad 6.80%
  543. 543 LFM2.5-8B-A1B val 6.70%
  544. 544 Phi-4-Reasoning-plus 6.42%
  545. 545 meta-llama/Llama-3.2-1B-Instruct 6.38%
  546. 546 Qwen3.5-2B-Base val 6.16%
  547. 547 HPLT/hplt2c_nob_checkpoints 5.74%
  548. 548 NbAiLab/borealis-270m val 5.49%
  549. 549 BSC-LT/salamandra-7b 5.14%
  550. 550 SmolLM2-360M 4.99%
  551. 551 NbAiLab/borealis-open-270m val 4.84%
  552. 552 Qwen3.5-0.8B-Base val 4.45%
  553. 553 openai-community/gpt2 3.67%
  554. 554 AI-Sweden-Models/gpt-sw3-356m-instruct 3.64%
  555. 555 SmolLM2-135M 3.55%
  556. 556 AI-Sweden-Models/gpt-sw3-6.7b-v2-instruct val 3.44%
  557. 557 baidu/ERNIE-4.5-0.3B-PT 3.29%
  558. 558 infly/OpenCoder-1.5B-Base 2.78%
  559. 559 AI-Sweden-Models/gpt-sw3-40b val 2.69%
  560. 560 openbmb/MiniCPM5-1B val 2.52%
  561. 561 openai-community/gpt2-medium 2.29%
  562. 562 NbAiLab/borealis-270m-instruct-preview val 2.13%
  563. 563 kyutai/helium-1-2b 1.77%
  564. 564 PleIAs/Baguettotron@refs/pr/6#no-thinking 1.74%
  565. 565 Muse-research/Muse-3B val 1.62%
  566. 566 openai-community/gpt2-large 1.44%
  567. 567 OpenLLM-France/Luciole-1B-Base val 1.16%
  568. 568 HPLT/hplt2c_dan_checkpoints 0.96%
  569. 569 HuggingFaceTB/SmolLM2-360M-Instruct 0.77%
  570. 570 AI-Sweden-Models/gpt-sw3-6.7b-v2-translator val 0.50%
  571. 571 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct 0.27%
  572. 572 facebook/MobileLLM-R1-360M 0.25%
  573. 573 ltg/norbert 0.15%
  574. 574 facebook/MobileLLM-R1-140M 0.12%
  575. 575 AI-Sweden-Models/gpt-sw3-1.3b val 0%
  576. 576 AI-Sweden-Models/gpt-sw3-20b 0%
  577. 577 AI-Sweden-Models/gpt-sw3-356m 0%
  578. 578 AI-Sweden-Models/gpt-sw3-6.7b val 0%
  579. 579 AI-Sweden-Models/gpt-sw3-6.7b-v2 val 0%
  580. 580 facebook/MobileLLM-R1-140M-base 0%
  581. 581 facebook/MobileLLM-R1-950M 0%
  582. 582 NorwAI/NorwAI-Magistral-24B-reasoning val 0%
  583. 583 openGPT-X/Teuken-7B-base-v0.6 val 0%
  584. 584 Qwen/Qwen2.5-Math-1.5B 0%
  585. 585 utter-project/EuroLLM-1.7B val 0%
  586. 586 PleIAs/Baguettotron@refs/pr/6#thinking -0.02%
  587. 587 facebook/MobileLLM-R1-360M-base -0.35%
  588. 588 AI-Sweden-Models/gpt-sw3-126m -0.37%
  589. 589 danish-foundation-models/munin-7b-open-it -0.47%
  590. 590 HPLT/hplt2c_nno_checkpoints -0.72%
  591. 591 inclusionAI/Ling-mini-base-2.0 val -0.89%
  592. 592 Shekswess/trlm-135m -0.98%
  593. 593 PleIAs/Baguettotron@refs/pr/6 -1.12%
  594. 594 inclusionAI/Ling-mini-2.0 val -1.15%
  595. 595 norallm/normistral-11b-translate val -1.60%
  596. 596 AIDC-AI/Marco-Mini-Global-Base val -2.03%

Source: Official Angry Tweets leaderboard

149 models. Scroll sideways for the rest.

  1. 67.0 GPT-6.2 Astra
  2. 66.3 GLM-5.3-Flash
  3. 65.3 GPT-6 Astra
  4. 64.2 Gemini 3.7 Flash
  5. 63.8 Gemini 2.5 Flash
  6. 63.5 Gemma 4 31B IT
  7. 62.8 GPT-6.2 Sol
  8. 62.6 GPT-6.1 Sol
  9. 62.2 Mistral Medium 3.5
  10. 62.2 o3-mini
  11. 61.9 Gemini 3.5 Flash
  12. 60.9 Qwen3.5 397B-A17B
  13. 60.6 Claude Opus 4.8
  14. 60.6 Qwen 3.8 27B
  15. 60.2 GPT-6 Luna
  16. 60.2 Gemini 3.6 Flash
  17. 60.2 Gemini 3 Pro
  18. 60.0 Olmo 3.1 32B Think
  19. 60.0 Gemma 4 12B
  20. 59.3 GPT-5.6 Sol
  21. 58.8 GPT-5
  22. 58.7 Reka Flash 3.1
  23. 58.6 Cogito v1 Preview 70B
  24. 58.5 GPT-5.2
  25. 58.5 glm-4.7-flash
  26. 58.4 Llama 3.1-70B
  27. 58.2 Qwen3.5-9B
  28. 58.1 Gemini 3.1 Flash-Lite
  29. 58.1 o3
  30. 58.1 Llama 4 Scout
  31. 58.0 Reka Flash 3
  32. 58.0 EXAONE 4.0 32B
  33. 57.7 Claude Sonnet 4.6
  34. 57.5 o4-mini
  35. 57.4 Gemma 3 27B
  36. 57.0 GPT-5.5
  37. 56.9 Gemma 4 26B A4B
  38. 56.9 GPT-4.1
  39. 56.9 Nemotron-Cascade 2 30B-A3B
  40. 56.8 Llama 3.3 70B Instruct
  41. 56.7 Gemma 3 12B
  42. 56.6 QwQ-32B
  43. 56.4 Apertus 1.5 70B
  44. 56.2 OLMo 2 32B Instruct
  45. 56.2 GLM-4.5-Air
  46. 56.1 Qwen3.5-35B-A3B-Base
  47. 56.0 Aya Expanse 8B
  48. 56.0 GPT-5 nano
  49. 55.9 Claude Haiku 4.5
  50. 55.9 Grok 4.20
  51. 55.7 Gemini 2.5 Pro
  52. 55.6 gpt-oss-120b
  53. 55.5 Qwen3-14B
  54. 55.5 Ministral 3 3B Instruct
  55. 55.5 GPT-5.6 Terra
  56. 55.4 Apertus 70B
  57. 55.4 Gemma 3n E4B
  58. 55.4 Qwen3-235B-A22B
  59. 55.2 Gemma 3n E2B
  60. 55.2 GPT-6 Sol
  61. 55.1 Gemini 3.5 Flash-Lite
  62. 55.1 Qwen3-Next-80B-A3B
  63. 55.0 Grok-2
  64. 55.0 Phi-4
  65. 54.8 Grok 3
  66. 54.6 Voxtral Small
  67. 54.5 Qwen3-8B
  68. 54.5 gpt-oss-20b
  69. 54.5 Grok-3 mini
  70. 54.3 Gemma 3 27B PT
  71. 54.1 Grok 4.1 Fast
  72. 54.1 Aya Expanse 32B
  73. 54.0 GLM-4.1V-9B-Thinking
  74. 53.3 Qwen3-Next-80B-A3B-Instruct
  75. 53.2 Gemini 2.5 Flash-Lite
  76. 53.0 Gemma 3 12B PT
  77. 52.8 Ministral 3 14B
  78. 52.8 DeepSeek-R1-Distill-Llama-70B
  79. 52.8 Gemma 4 E4B
  80. 52.7 GPT-5 mini
  81. 52.6 Llama-3.1-Tulu-3-8B
  82. 52.5 GPT-5.6 Luna
  83. 52.3 Olmo 3.1 32B Instruct
  84. 52.1 Command R7B
  85. 52.1 Qwen3-4B-Thinking-2507
  86. 51.9 Llama 3.1-405B
  87. 51.8 Granite 4.0 H Small
  88. 51.1 Trinity Mini
  89. 51.0 Qwen1.5-14B-Chat
  90. 50.8 Granite-3.0-8B-Instruct
  91. 50.8 Qwen3.5-4B
  92. 49.5 Gemma 4 E2B
  93. 49.5 DeepSeek-R1-Distill-Qwen-32B
  94. 49.4 Ministral 3 8B Instruct
  95. 49.3 Qwen3-4B
  96. 49.1 Falcon3-10B-Instruct
  97. 49.0 deepseek-r1-0528-qwen3-8b
  98. 49.0 Gemma 2 27B
  99. 48.7 Ministral 3 3B Reasoning
  100. 48.6 Tiny Aya Water
  101. 48.1 Granite-3.0-8B-Base
  102. 48.0 Llama 3.1-8B
  103. 48.0 nvidia-nemotron-nano-9b-v2
  104. 48.0 granite-4.0-micro
  105. 47.9 Tiny Aya Global
  106. 47.8 Falcon3-7B-Instruct
  107. 47.4 qwen3-4b-instruct-2507
  108. 47.4 Mistral 7B v0.3
  109. 47.3 Gemma 7B
  110. 47.1 Gemma 2 2B IT
  111. 46.6 Yi-1.5-34B
  112. 45.9 DeepSeek-R1-Distill-Qwen-14B
  113. 45.8 Apertus 1.5 8B
  114. 45.7 Qwen3-1.7B
  115. 45.2 Mistral 7B v0.1
  116. 44.9 Mistral-7B-Instruct-v0.2
  117. 44.7 Mistral-7B-Instruct-v0.1
  118. 43.4 Qwen1.5-14B
  119. 42.7 Qwen3.5-2B
  120. 42.5 AFM-4.5B
  121. 41.9 Gemma 3 1B
  122. 41.0 Qwen3.5-0.8B
  123. 40.4 Phi-4-reasoning
  124. 39.3 Gemma 2 9B
  125. 39.2 Apertus 8B Instruct
  126. 38.9 Gemma 2B
  127. 36.7 Gemma 3 1B PT
  128. 36.1 Hunyuan-4B-Instruct
  129. 30.8 Falcon3-3B-Instruct
  130. 28.0 Llama 3.2 1B
  131. 27.4 Qwen3-0.6B
  132. 25.8 LFM2.5-1.2B-Thinking
  133. 24.3 Qwen3.5-9B-Base
  134. 23.2 LFM2.5-1.2B-Instruct
  135. 20.0 DeepSeek-R1-Distill-Qwen-7B
  136. 19.4 Phi-2
  137. 19.2 Phi-4-mini-reasoning
  138. 17.4 DeepSeek-R1-Distill-Qwen-1.5B
  139. 16.9 LFM2-1.2B
  140. 16.2 Gemma 3 270M
  141. 15.7 SmolLM2 1.7B Instruct
  142. 8.53 SmolLM2-1.7B
  143. 6.93 Qwen3.5-4B-Base
  144. 6.70 LFM2.5-8B-A1B
  145. 6.42 Phi-4-Reasoning-plus
  146. 6.16 Qwen3.5-2B-Base
  147. 4.99 SmolLM2-360M
  148. 4.45 Qwen3.5-0.8B-Base
  149. 3.55 SmolLM2-135M
  1. 1 GPT-6.2 Astra zero-shot, val 67%
  2. 2 GLM-5.3-Flash val 66.3%
  3. 3 GPT-6 Astra zero-shot, val 65.3%
  4. 4 Gemini 3.7 Flash zero-shot, val 64.2%
  5. 5 Gemini 2.5 Flash Sep 2025, zero-shot, val 63.8%
  6. 6 Gemma 4 31B IT val 63.5%
  7. 7 GPT-6.2 Sol zero-shot, val 62.8%
  8. 8 GPT-6.1 Sol zero-shot, val 62.6%
  9. 9 Mistral Medium 3.5 val 62.2%
  10. 10 o3-mini zero-shot, val 62.2%
  11. 11 Gemini 3.5 Flash zero-shot, val 61.9%
  12. 12 Qwen3.5 397B-A17B 60.9%
  13. 13 Claude Opus 4.8 zero-shot, val 60.6%
  14. 14 Qwen 3.8 27B val 60.6%
  15. 15 GPT-6 Luna zero-shot, val 60.2%
  16. 16 Gemini 3.6 Flash zero-shot, val 60.2%
  17. 17 Gemini 3 Pro zero-shot, val 60.2%
  18. 18 Olmo 3.1 32B Think val 60%
  19. 19 Gemma 4 12B val 60%
  20. 20 GPT-5.6 Sol zero-shot, val 59.3%
  21. 21 GPT-5 zero-shot, val 58.8%
  22. 22 Reka Flash 3.1 58.7%
  23. 23 Cogito v1 Preview 70B 58.6%
  24. 24 GPT-5.2 zero-shot, val 58.5%
  25. 25 glm-4.7-flash val 58.5%
  26. 26 Llama 3.1-70B 58.4%
  27. 27 Qwen3.5-9B val 58.2%
  28. 28 Gemini 3.1 Flash-Lite zero-shot, val 58.1%
  29. 29 o3 zero-shot, val 58.1%
  30. 30 Llama 4 Scout 58.1%
  31. 31 Reka Flash 3 58%
  32. 32 EXAONE 4.0 32B val 58%
  33. 33 Claude Sonnet 4.6 zero-shot, val 57.7%
  34. 34 o4-mini zero-shot, val 57.5%
  35. 35 Gemma 3 27B 57.4%
  36. 36 GPT-5.5 zero-shot, val 57%
  37. 37 Gemma 4 26B A4B val 56.9%
  38. 38 GPT-4.1 zero-shot, val 56.9%
  39. 39 Nemotron-Cascade 2 30B-A3B val 56.9%
  40. 40 Llama 3.3 70B Instruct zero-shot, val 56.8%
  41. 41 Gemma 3 12B 56.7%
  42. 42 QwQ-32B 56.6%
  43. 43 Apertus 1.5 70B val 56.4%
  44. 44 OLMo 2 32B Instruct 56.2%
  45. 45 GLM-4.5-Air thinking 56.2%
  46. 46 Qwen3.5-35B-A3B-Base val 56.1%
  47. 47 Aya Expanse 8B val 56%
  48. 48 GPT-5 nano zero-shot, val 56%
  49. 49 Claude Haiku 4.5 zero-shot, val 55.9%
  50. 50 Grok 4.20 thinking, zero-shot, val 55.9%
  51. 51 Gemini 2.5 Pro Jun 2025, zero-shot, val 55.7%
  52. 52 gpt-oss-120b val 55.6%
  53. 53 Qwen3-14B 55.5%
  54. 54 Ministral 3 3B Instruct zero-shot, val 55.5%
  55. 55 GPT-5.6 Terra zero-shot, val 55.5%
  56. 56 Apertus 70B 55.4%
  57. 57 Gemma 3n E4B 55.4%
  58. 58 Qwen3-235B-A22B 55.4%
  59. 59 Gemma 3n E2B 55.2%
  60. 60 GPT-6 Sol zero-shot, val 55.2%
  61. 61 Gemini 3.5 Flash-Lite zero-shot, val 55.1%
  62. 62 Qwen3-Next-80B-A3B thinking 55.1%
  63. 63 Grok-2 Dec 2024, zero-shot, val 55%
  64. 64 Phi-4 55%
  65. 65 Grok 3 zero-shot, val 54.8%
  66. 66 Voxtral Small 54.6%
  67. 67 Qwen3-8B 54.5%
  68. 68 gpt-oss-20b high 54.5%
  69. 69 Grok-3 mini zero-shot, val 54.5%
  70. 70 Gemma 3 27B PT 54.3%
  71. 71 Grok 4.1 Fast zero-shot, val 54.1%
  72. 72 Aya Expanse 32B 54.1%
  73. 73 GLM-4.1V-9B-Thinking 54%
  74. 74 Qwen3-Next-80B-A3B-Instruct 53.3%
  75. 75 Gemini 2.5 Flash-Lite Sep 2025, zero-shot, val 53.2%
  76. 76 Gemma 3 12B PT 53%
  77. 77 Ministral 3 14B val 52.8%
  78. 78 DeepSeek-R1-Distill-Llama-70B 52.8%
  79. 79 Gemma 4 E4B val 52.8%
  80. 80 GPT-5 mini zero-shot, val 52.7%
  81. 81 Llama-3.1-Tulu-3-8B 52.6%
  82. 82 GPT-5.6 Luna zero-shot, val 52.5%
  83. 83 Olmo 3.1 32B Instruct 52.3%
  84. 84 Command R7B 52.1%
  85. 85 Qwen3-4B-Thinking-2507 52.1%
  86. 86 Llama 3.1-405B 51.9%
  87. 87 Granite 4.0 H Small 51.8%
  88. 88 Trinity Mini 51.1%
  89. 89 Qwen1.5-14B-Chat 51%
  90. 90 Granite-3.0-8B-Instruct 50.8%
  91. 91 Qwen3.5-4B val 50.8%
  92. 92 Gemma 4 E2B val 49.5%
  93. 93 DeepSeek-R1-Distill-Qwen-32B 49.5%
  94. 94 Ministral 3 8B Instruct val 49.4%
  95. 95 Qwen3-4B 49.3%
  96. 96 Falcon3-10B-Instruct 49.1%
  97. 97 deepseek-r1-0528-qwen3-8b 49%
  98. 98 Gemma 2 27B 49%
  99. 99 Ministral 3 3B Reasoning 48.7%
  100. 100 Tiny Aya Water val 48.6%
  101. 101 Granite-3.0-8B-Base 48.1%
  102. 102 Llama 3.1-8B 48%
  103. 103 nvidia-nemotron-nano-9b-v2 48%
  104. 104 granite-4.0-micro 48%
  105. 105 Tiny Aya Global val 47.9%
  106. 106 Falcon3-7B-Instruct 47.8%
  107. 107 qwen3-4b-instruct-2507 47.4%
  108. 108 Mistral 7B v0.3 47.4%
  109. 109 Gemma 7B 47.3%
  110. 110 Gemma 2 2B IT 47.1%
  111. 111 Yi-1.5-34B 46.6%
  112. 112 DeepSeek-R1-Distill-Qwen-14B 45.9%
  113. 113 Apertus 1.5 8B zero-shot, val 45.8%
  114. 114 Qwen3-1.7B 45.7%
  115. 115 Mistral 7B v0.1 45.2%
  116. 116 Mistral-7B-Instruct-v0.2 44.9%
  117. 117 Mistral-7B-Instruct-v0.1 44.7%
  118. 118 Qwen1.5-14B 43.4%
  119. 119 Qwen3.5-2B val 42.7%
  120. 120 AFM-4.5B 42.5%
  121. 121 Gemma 3 1B 41.9%
  122. 122 Qwen3.5-0.8B val 41%
  123. 123 Phi-4-reasoning 40.4%
  124. 124 Gemma 2 9B 39.3%
  125. 125 Apertus 8B Instruct 39.2%
  126. 126 Gemma 2B 38.9%
  127. 127 Gemma 3 1B PT 36.7%
  128. 128 Hunyuan-4B-Instruct 36.1%
  129. 129 Falcon3-3B-Instruct 30.8%
  130. 130 Llama 3.2 1B 28%
  131. 131 Qwen3-0.6B 27.4%
  132. 132 LFM2.5-1.2B-Thinking 25.8%
  133. 133 Qwen3.5-9B-Base val 24.3%
  134. 134 LFM2.5-1.2B-Instruct 23.2%
  135. 135 DeepSeek-R1-Distill-Qwen-7B 20%
  136. 136 Phi-2 19.4%
  137. 137 Phi-4-mini-reasoning 19.2%
  138. 138 DeepSeek-R1-Distill-Qwen-1.5B 17.4%
  139. 139 LFM2-1.2B 16.9%
  140. 140 Gemma 3 270M 16.2%
  141. 141 SmolLM2 1.7B Instruct 15.7%
  142. 142 SmolLM2-1.7B 8.53%
  143. 143 Qwen3.5-4B-Base val 6.93%
  144. 144 LFM2.5-8B-A1B val 6.70%
  145. 145 Phi-4-Reasoning-plus 6.42%
  146. 146 Qwen3.5-2B-Base val 6.16%
  147. 147 SmolLM2-360M 4.99%
  148. 148 Qwen3.5-0.8B-Base val 4.45%
  149. 149 SmolLM2-135M 3.55%

Source: Official Angry Tweets leaderboard

#EntryScore ±SettingDateSource
1 GPT-6.2 Astra 67% ±1.1 zero-shot, val Nov 2026 Official leaderboard
2 GLM-5.3-Flash 66.3% ±0.8 val Aug 2026 Official leaderboard
3 GPT-6 Astra 65.3% ±1.3 zero-shot, val Sep 2026 Official leaderboard
4 mistralai/Mistral-Large-3-675B-Instruct-2512 64.5% ±1.1 val Official leaderboard
5 claude-sonnet-4-5-20250929#thinking 64.3% ±0.9 zero-shot, val Official leaderboard
6 Gemini 3.7 Flash 64.2% ±0.9 zero-shot, val Aug 2026 Official leaderboard
7 danish-foundation-models/DFM-Mimir-v1.5 64% ±0.8 val Official leaderboard
8 Gemini 2.5 Flash Sep 2025 63.8% ±0.9 zero-shot, val Sep 2025 Official leaderboard
9 speakleash/Bielik-11B-v3.1-Instruct 63.6% val Official leaderboard
10 Qwen/Qwen3.5-27B-FP8 63.5% ±1.5 val Official leaderboard
11 Gemma 4 31B IT 63.5% ±1 val Apr 2026 Official leaderboard
12 NbAiLab/borealis2-31b-preview 63.3% ±1.2 val Official leaderboard
13 claude-3-7-sonnet-20250219#no-thinking 63.2% ±0.7 zero-shot, val Official leaderboard
14 Qwen/Qwen3.6-27B-FP8 63.2% ±0.8 val Official leaderboard
15 syvai/danskgpt-v4-31b 63.2% ±0.6 zero-shot, val Official leaderboard
16 deepseek/deepseek-flash#no-thinking 63.2% ±0.8 val Official leaderboard
17 GPT-6.2 Sol 62.8% ±1.3 zero-shot, val Nov 2026 Official leaderboard
18 speakleash/Bielik-11B-v3.0-Instruct 62.6% val Official leaderboard
19 claude-sonnet-4-5-20250929#no-thinking 62.6% ±1.1 zero-shot, val Official leaderboard
20 GPT-6.1 Sol 62.6% ±1 zero-shot, val Sep 2026 Official leaderboard
21 Mistral Medium 3.5 62.2% ±1.1 val Apr 2026 Official leaderboard
22 o3-mini 62.2% ±0.9 zero-shot, val Jan 2025 Official leaderboard
23 Gemini 3.5 Flash 61.9% ±1 zero-shot, val May 2026 Official leaderboard
24 gemini/gemini-3-flash-preview#thinking 61.7% ±0.9 zero-shot, val Official leaderboard
25 zai-org/GLM-4.7-FP8 61.6% ±0.7 val Official leaderboard
26 danish-foundation-models/DFM-Mimir-v1.8 61.5% ±0.6 val Official leaderboard
27 Qwen3.5 397B-A17B 60.9% ±0.6 Feb 2026 Official leaderboard
28 meta-llama/Llama-3.1-405B-Instruct-FP8 60.8% ±0.5 Official leaderboard
29 Claude Opus 4.8 60.6% ±0.8 zero-shot, val May 2026 Official leaderboard
30 deepseek/deepseek-flash#low 60.6% ±1.1 zero-shot, val Official leaderboard
31 Qwen 3.8 27B 60.6% ±1.1 val Aug 2026 Official leaderboard
32 Qwen/Qwen3.5-35B-A3B-FP8 60.5% ±1.1 val Official leaderboard
33 GPT-6 Luna 60.2% ±1.3 zero-shot, val Sep 2026 Official leaderboard
34 Gemini 3.6 Flash 60.2% ±0.9 zero-shot, val Jul 2026 Official leaderboard
35 Gemini 3 Pro 60.2% ±0.7 zero-shot, val Nov 2025 Official leaderboard
36 Olmo 3.1 32B Think 60% ±1.3 val Dec 2025 Official leaderboard
37 Gemma 4 12B 60% ±0.9 val Jun 2026 Official leaderboard
38 NbAiLab/borealis-open-27b 59.8% ±0.8 val Official leaderboard
39 danish-foundation-models/DFM-Mimir 59.6% ±0.8 val Official leaderboard
40 NbAiLab/borealis-27b-instruct-preview 59.5% ±0.9 val Official leaderboard
41 GPT-5.6 Sol 59.3% ±1.3 zero-shot, val Jul 2026 Official leaderboard
42 Qwen/Qwen3-235B-A22B-Thinking-2507-FP8 59.3% ±0.4 Official leaderboard
43 LumiOpen/Llama-Poro-2-70B-SFT 59.1% ±0.4 Official leaderboard
44 mistralai/Mistral-Small-3.1-24B-Instruct-2503 59.1% ±1.2 zero-shot, val Official leaderboard
45 ThatsGroes/gemma-2-27b-it-FP8-Dynamic 58.9% ±0.8 Official leaderboard
46 GPT-5 58.8% ±1 zero-shot, val Aug 2025 Official leaderboard
47 deepcogito/cogito-v2-preview-llama-70B 58.8% ±0.6 Official leaderboard
48 gemini/gemini-3-flash-preview#no-thinking 58.7% ±1 zero-shot, val Official leaderboard
49 Reka Flash 3.1 58.7% ±0.6 May 2025 Official leaderboard
50 NbAiLab/borealis-27b 58.6% ±1.3 val Official leaderboard
51 Cogito v1 Preview 70B 58.6% ±0.7 Apr 2025 Official leaderboard
52 LumiOpen/Llama-Poro-2-70B-Instruct 58.5% ±0.5 Official leaderboard
53 GPT-5.2 58.5% ±0.6 zero-shot, val Dec 2025 Official leaderboard
54 glm-4.7-flash 58.5% ±1.2 val Jan 2026 Official leaderboard
55 Llama 3.1-70B 58.4% ±0.9 Jul 2024 Official leaderboard
56 claude-3-7-sonnet-20250219#thinking 58.3% ±1 zero-shot, val Official leaderboard
57 mistralai/Mistral-Small-4-119B-2603 58.3% ±0.9 val Official leaderboard
58 Qwen3.5-9B 58.2% ±1.9 val Feb 2026 Official leaderboard
59 GadflyII/GLM-4.7-Flash-NVFP4 58.2% ±1.2 val Official leaderboard
60 openai/gpt-5.4-mini-2026-03-17#medium 58.1% ±1.2 zero-shot, val Official leaderboard
61 Gemini 3.1 Flash-Lite 58.1% ±0.9 zero-shot, val Mar 2026 Official leaderboard
62 o3 58.1% ±1 zero-shot, val Apr 2025 Official leaderboard
63 Llama 4 Scout 58.1% ±0.4 Apr 2025 Official leaderboard
64 Reka Flash 3 58% ±0.7 Mar 2025 Official leaderboard
65 EXAONE 4.0 32B 58% ±1.4 val Jul 2025 Official leaderboard
66 syvai/danskgpt-v4-12b 57.9% ±0.7 zero-shot, val Official leaderboard
67 Claude Sonnet 4.6 57.7% ±0.9 zero-shot, val Feb 2026 Official leaderboard
68 google/gemma-2-27b-it 57.6% ±0.9 Official leaderboard
69 o4-mini 57.5% ±0.6 zero-shot, val Apr 2025 Official leaderboard
70 Gemma 3 27B 57.4% ±0.8 Mar 2025 Official leaderboard
71 mistralai/Mistral-Small-24B-Instruct-2501 57.3% ±0.7 Official leaderboard
72 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 57.1% ±1.3 val Official leaderboard
73 GPT-5.5 57% ±1.4 zero-shot, val Apr 2026 Official leaderboard
74 speakleash/Bielik-11B-v2.3-Instruct 57% ±1 val Official leaderboard
75 typesafe/jev-1.13.0 57% ±0.6 zero-shot Official leaderboard
76 Gemma 4 26B A4B 56.9% ±0.8 val Apr 2026 Official leaderboard
77 GPT-4.1 56.9% ±1 zero-shot, val Apr 2025 Official leaderboard
78 Nemotron-Cascade 2 30B-A3B 56.9% ±1.2 val Mar 2026 Official leaderboard
79 Llama 3.3 70B Instruct 56.8% ±0.7 zero-shot, val Dec 2024 Official leaderboard
80 Unbabel/Tower-Plus-72B 56.8% ±1.2 val Official leaderboard
81 mistralai/Mistral-Small-3.2-24B-Instruct-2506 56.8% ±0.7 Official leaderboard
82 gpt-5-nano-2025-08-07#high 56.7% ±1.1 zero-shot, val Official leaderboard
83 mistralai/Mistral-Small-3.1-24B-Base-2503 56.7% ±0.9 Official leaderboard
84 Gemma 3 12B 56.7% ±0.7 Mar 2025 Official leaderboard
85 QwQ-32B 56.6% ±0.4 Mar 2025 Official leaderboard
86 ollama_chat/gemma3:27b-it-fp16 56.4% ±0.8 Official leaderboard
87 Apertus 1.5 70B 56.4% ±1.3 val Jul 2026 Official leaderboard
88 NbAiLab/borealis-12b-instruct-preview 56.3% ±1.1 val Official leaderboard
89 OLMo 2 32B Instruct 56.2% ±0.9 Mar 2025 Official leaderboard
90 danish-foundation-models/munin-qwen3.5-9B 56.2% ±0.6 Official leaderboard
91 GLM-4.5-Air 56.2% ±0.6 thinking Jul 2025 Official leaderboard
92 gpt-5-2025-08-07#high 56.2% ±0.9 zero-shot, val Official leaderboard
93 ollama_chat/gemma3:12b-it-q4_K_M 56.2% ±1.4 val Official leaderboard
94 deepcogito/cogito-v1-preview-qwen-32B 56.1% ±0.6 Official leaderboard
95 Qwen3.5-35B-A3B-Base 56.1% ±1.2 val Feb 2026 Official leaderboard
96 NbAiLab/borealis2-26b-a4b-preview 56.1% ±1.1 val Official leaderboard
97 ollama_chat/gemma3:27b-it-q4_K_M 56% ±1.1 val Official leaderboard
98 Aya Expanse 8B 56% ±1.4 val Oct 2024 Official leaderboard
99 GPT-5 nano 56% ±0.9 zero-shot, val Aug 2025 Official leaderboard
100 mistralai/Magistral-Small-2509 56% ±1.1 Official leaderboard
101 Qwen/Qwen3-235B-A22B-Instruct-2507-FP8 55.9% ±0.8 Official leaderboard
102 Claude Haiku 4.5 55.9% ±0.8 zero-shot, val Oct 2025 Official leaderboard
103 Grok 4.20 55.9% ±0.8 thinking, zero-shot, val Feb 2026 Official leaderboard
104 openai/gpt-5.4-2026-03-05#none 55.7% ±1.1 zero-shot, val Official leaderboard
105 NovaSky-AI/Sky-T1-32B-Preview 55.7% ±0.6 Official leaderboard
106 Qwen/Qwen3.5-9B#no-thinking 55.7% ±0.6 Official leaderboard
107 Gemini 2.5 Pro Jun 2025 55.7% ±0.9 zero-shot, val Jun 2025 Official leaderboard
108 zai-org/GLM-4.5-Air#no-thinking 55.6% ±0.9 Official leaderboard
109 gpt-oss-120b 55.6% ±0.8 val Aug 2025 Official leaderboard
110 Qwen3-14B 55.5% ±0.8 Apr 2025 Official leaderboard
111 NbAiLab/borealis2-e4b-preview 55.5% ±1.8 val Official leaderboard
112 Ministral 3 3B Instruct 55.5% ±0.8 zero-shot, val Dec 2025 Official leaderboard
113 GPT-5.6 Terra 55.5% ±1.4 zero-shot, val Jul 2026 Official leaderboard
114 Apertus 70B 55.4% ±0.7 Sep 2025 Official leaderboard
115 Gemma 3n E4B 55.4% ±1 Jun 2025 Official leaderboard
116 Qwen3-235B-A22B 55.4% ±0.6 Apr 2025 Official leaderboard
117 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8 55.3% ±1.3 val Official leaderboard
118 Qwen/Qwen3-14B#no-thinking 55.2% ±0.8 Official leaderboard
119 Aleph-Alpha/Kolibri-1 55.2% ±0.9 val Official leaderboard
120 meta-llama/Llama-3.1-8B-Instruct 55.2% ±0.7 Official leaderboard
121 Gemma 3n E2B 55.2% ±0.7 Jun 2025 Official leaderboard
122 GPT-6 Sol 55.2% ±1.4 zero-shot, val Sep 2026 Official leaderboard
123 Gemini 3.5 Flash-Lite 55.1% ±1.1 zero-shot, val Jul 2026 Official leaderboard
124 Qwen3-Next-80B-A3B 55.1% ±0.8 thinking Sep 2025 Official leaderboard
125 openai/gpt-5.4-mini-2026-03-17#high 55.1% ±1 zero-shot, val Official leaderboard
126 NbAiLab/borealis-open-12b 55% ±0.9 val Official leaderboard
127 Grok-2 Dec 2024 55% ±1.5 zero-shot, val Dec 2024 Official leaderboard
128 Phi-4 55% ±0.7 Dec 2024 Official leaderboard
129 deepcogito/cogito-v1-preview-qwen-14B 54.9% ±1 Official leaderboard
130 Grok 3 54.8% ±1 zero-shot, val Apr 2025 Official leaderboard
131 speakleash/Bielik-Minitron-7B-v3.0-Instruct 54.7% val Official leaderboard
132 Qwen/Qwen3-32B#no-thinking 54.6% ±1 Official leaderboard
133 Voxtral Small 54.6% ±0.9 Jul 2025 Official leaderboard
134 Qwen3-8B 54.5% ±0.9 Apr 2025 Official leaderboard
135 gpt-oss-20b 54.5% ±0.8 high Aug 2025 Official leaderboard
136 Grok-3 mini 54.5% ±1 zero-shot, val Apr 2025 Official leaderboard
137 NousResearch/DeepHermes-3-Llama-3-8B-Preview 54.4% ±0.7 Official leaderboard
138 utter-project/EuroLLM-9B-Instruct 54.4% ±0.6 Official leaderboard
139 Gemma 3 27B PT 54.3% ±0.9 Mar 2025 Official leaderboard
140 deepcogito/cogito-v1-preview-llama-8B 54.2% ±1 Official leaderboard
141 NbAiLab/borealis-12b 54.2% ±1.1 val Official leaderboard
142 timpal0l/BeagleCatMunin 54.1% ±1.1 val Official leaderboard
143 Grok 4.1 Fast 54.1% ±1 zero-shot, val Nov 2025 Official leaderboard
144 Aya Expanse 32B 54.1% ±0.7 Oct 2024 Official leaderboard
145 alexandrainst/Meta-Llama-3.1-8B-laerebogen-and-dialog-k2-da 54% ±1.4 val Official leaderboard
146 GLM-4.1V-9B-Thinking 54% ±0.7 Jul 2025 Official leaderboard
147 openai/gpt-5.4-mini-2026-03-17#low 53.9% ±0.9 zero-shot, val Official leaderboard
148 ollama_chat/gemma3:12b-it-fp16 53.6% ±1.5 val Official leaderboard
149 gpt-5-2025-08-07#minimal 53.6% ±1 zero-shot, val Official leaderboard
150 alexandrainst/Meta-Llama-3.1-8B-laerebogen 53.4% ±1.2 val Official leaderboard
151 gemini/gemini-2.5-flash#no-thinking 53.4% ±0.8 zero-shot, val Official leaderboard
152 Qwen3-Next-80B-A3B-Instruct 53.3% ±0.7 Sep 2025 Official leaderboard
153 ollama_chat/phi4:14b-q4_K_M 53.3% ±1.4 val Official leaderboard
154 allenai/Olmo-3.1-32B-Instruct-DPO 53.2% ±0.9 Official leaderboard
155 KennethEnevoldsen/dfm-sentence-encoder-large 53.2% ±0.8 Official leaderboard
156 gemini/gemini-2.5-flash-lite-preview-06-17#thinking 53.2% ±1.2 zero-shot, val Official leaderboard
157 Gemini 2.5 Flash-Lite Sep 2025 53.2% ±1.2 zero-shot, val Sep 2025 Official leaderboard
158 google/gemma-2-9b-it 53.2% ±0.6 Official leaderboard
159 utter-project/EuroLLM-22B-Instruct-2512 53.1% ±0.4 Official leaderboard
160 gpt-5-mini-2025-08-07#high 53% ±0.8 zero-shot, val Official leaderboard
161 Gemma 3 12B PT 53% ±0.7 Mar 2025 Official leaderboard
162 utter-project/EuroLLM-9B 53% ±1.1 Official leaderboard
163 allenai/Llama-3.1-Tulu-3-8B-DPO 52.9% ±0.8 Official leaderboard
164 intfloat/multilingual-e5-large 52.9% ±0.9 Official leaderboard
165 mistralai/Ministral-3-14B-Base-2512 52.9% ±1.3 val Official leaderboard
166 allenai/Olmo-3.1-32B-Instruct-SFT 52.8% ±0.8 Official leaderboard
167 Ministral 3 14B 52.8% ±1 val Dec 2025 Official leaderboard
168 danish-foundation-models/munin-ministral3-8B 52.8% ±0.7 Official leaderboard
169 DeepSeek-R1-Distill-Llama-70B 52.8% ±0.6 Jan 2025 Official leaderboard
170 Gemma 4 E4B 52.8% ±1.2 val Mar 2026 Official leaderboard
171 GPT-5 mini 52.7% ±0.9 zero-shot, val Aug 2025 Official leaderboard
172 ordbogen/odin-medium 52.7% ±0.8 zero-shot, val Official leaderboard
173 Qwen/Qwen3-30B-A3B#no-thinking 52.7% ±1 Official leaderboard
174 timpal0l/dolphin-2.9-llama3-8b-flashback 52.6% ±1.3 val Official leaderboard
175 Llama-3.1-Tulu-3-8B 52.6% ±1 Nov 2024 Official leaderboard
176 GPT-5.6 Luna 52.5% ±1.1 zero-shot, val Jul 2026 Official leaderboard
177 merge-crew/da-sv-ties 52.5% ±1.5 val Official leaderboard
178 utter-project/EuroLLM-22B-Instruct-Preview 52.5% ±0.6 Official leaderboard
179 Olmo 3.1 32B Instruct 52.3% ±0.9 Dec 2025 Official leaderboard
180 alexandrainst/Meta-Llama-3.1-8B-laerebogen-two-epochs 52.3% ±1.6 val Official leaderboard
181 01-ai/Yi-1.5-9B 52.3% ±0.7 Official leaderboard
182 birgermoell/BeagleCatMunin-Flashback-Bellman 52.2% ±1.1 val Official leaderboard
183 mistralai/Ministral-3-14B-Instruct-2512 52.2% ±1.2 val Official leaderboard
184 Command R7B 52.1% ±0.5 Dec 2024 Official leaderboard
185 Qwen3-4B-Thinking-2507 52.1% ±0.8 Aug 2025 Official leaderboard
186 ordbogen/odin-large 52.1% ±1 zero-shot, val Official leaderboard
187 RJuro/munin-neuralbeagle-7b 52% ±1.4 val Official leaderboard
188 meta-llama/Llama-4-Scout-17B-16E 52% ±0.8 Official leaderboard
189 speakleash/Bielik-11B-v2.6-Instruct 52% ±0.8 Official leaderboard
190 bineric/lynx-instruct-30b 51.9% ±1.1 val Official leaderboard
191 Llama 3.1-405B 51.9% ±0.7 Jul 2024 Official leaderboard
192 merge-crew/da-sv-dare-ties-density-0.9 51.9% ±1.8 val Official leaderboard
193 openGPT-X/Teuken-7B-instruct-v0.6 51.8% ±1.1 val Official leaderboard
194 merge-crew/da-sv-task-arithmetic 51.8% ±1.8 val Official leaderboard
195 Granite 4.0 H Small 51.8% ±1 Oct 2025 Official leaderboard
196 birgermoell/Flashback-Bellman 51.6% ±1 val Official leaderboard
197 birgermoell/Rapid-Cycling 51.6% ±1 val Official leaderboard
198 TrustLLMeu/baseline-7-8b_2-3t-tokens_llama 51.6% ±0.9 Official leaderboard
199 Qwen/Qwen3-8B#no-thinking 51.6% ±1.3 Official leaderboard
200 openai/gpt-5.4-mini-2026-03-17#none 51.6% ±1.2 zero-shot, val Official leaderboard
201 merge-crew/da-sv-slerp 51.5% ±1.5 val Official leaderboard
202 gemini/gemini-2.5-flash-lite-preview-06-17#no-thinking 51.5% ±1.1 zero-shot, val Official leaderboard
203 gemini/gemini-2.5-flash-lite#no-thinking 51.5% ±1.1 zero-shot, val Official leaderboard
204 CohereForAI/aya-23-35B 51.3% ±1.1 Official leaderboard
205 openai/gpt-5.4-nano-2026-03-17#medium 51.3% ±1.4 zero-shot, val Official leaderboard
206 RJuro/munin-neuralbeagle-SkoleGPTOpenOrca-7b 51.1% ±1.3 val Official leaderboard
207 LumiOpen/Viking-33B 51.1% ±0.7 Official leaderboard
208 bullerwins/Magistral-Small-2506-fp8 51.1% ±0.7 Official leaderboard
209 Trinity Mini 51.1% ±0.6 Dec 2025 Official leaderboard
210 Qwen1.5-14B-Chat 51% ±0.7 Feb 2024 Official leaderboard
211 danish-foundation-models/Meta-Llama-3.1-8B-laerebogen 50.9% ±1 Official leaderboard
212 tiiuae/falcon-11B 50.9% ±0.7 Official leaderboard
213 Granite-3.0-8B-Instruct 50.8% ±1.1 Oct 2024 Official leaderboard
214 Qwen3.5-4B 50.8% ±1.8 val Feb 2026 Official leaderboard
215 NbAiLab/nb-roberta-base-scandi 50.7% ±1 Official leaderboard
216 openai/gpt-5.4-nano-2026-03-17#low 50.7% ±0.8 zero-shot, val Official leaderboard
217 danish-foundation-models/DFM-Mimir-v1.8-Mini 50.7% ±0.6 zero-shot, val Official leaderboard
218 Unbabel/Tower-Plus-9B 50.6% ±1.5 val Official leaderboard
219 allenai/Llama-3.1-Tulu-3.1-8B 50.5% ±1.2 Official leaderboard
220 alexandrainst/Meta-Llama-3.1-8B-laerebogen-and-filtered-dialog-k2-da 50.5% ±1.3 val Official leaderboard
221 meta-llama/Llama-2-70b-hf 50.5% ±1 Official leaderboard
222 ollama_chat/hf.co/google/gemma-3-4b-it-qat-q4_0-gguf 50.5% ±0.8 Official leaderboard
223 ollama_chat/gemma3:4b-it-q4_K_M 50.3% ±0.7 Official leaderboard
224 openai/gpt-5.4-nano-2026-03-17#none 50.3% ±1 zero-shot, val Official leaderboard
225 senseable/WestLake-7B-v2 50.3% ±0.8 Official leaderboard
226 danish-foundation-models/encoder-large-v1 50.3% ±0.9 Official leaderboard
227 merge-crew/da-sv-dare-ties-density-0.6 50.3% ±1.5 val Official leaderboard
228 152334H/miqu-1-70b-sf 50.2% ±0.8 Official leaderboard
229 mistralai/Ministral-3-8B-Reasoning-2512 50.1% ±0.9 Official leaderboard
230 deepcogito/cogito-v1-preview-llama-3B 50% ±0.9 Official leaderboard
231 KennethEnevoldsen/dfm-sentence-encoder-medium 50% ±0.8 Official leaderboard
232 intfloat/multilingual-e5-base 49.8% ±0.6 Official leaderboard
233 meta-llama/Llama-2-70b-chat-hf 49.8% ±0.8 Official leaderboard
234 NbAiLab/nb-roberta-base-scandi-1e4 49.7% ±1 Official leaderboard
235 AI-Sweden-Models/Llama-3-8B-instruct 49.7% ±1.3 val Official leaderboard
236 NbAiLab/nb-roberta-base-ncc-plus-scandi-1e4 49.7% ±1.5 Official leaderboard
237 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-Base-BF16 49.7% ±1.6 val Official leaderboard
238 AI-Sweden-Models/roberta-large-1160k 49.6% ±1.4 Official leaderboard
239 AI-Sweden-Models/tyr 49.6% ±1.5 val Official leaderboard
240 openai/gpt-5.4-nano-2026-03-17#high 49.6% ±0.9 zero-shot, val Official leaderboard
241 Gemma 4 E2B 49.5% ±1.2 val Mar 2026 Official leaderboard
242 NbAiLab/borealis-4b 49.5% ±1.1 val Official leaderboard
243 utter-project/EuroLLM-22B-Preview 49.5% ±1.1 Official leaderboard
244 DeepSeek-R1-Distill-Qwen-32B 49.5% ±0.8 Jan 2025 Official leaderboard
245 NbAiLab/nb-roberta-base-ncc-plus-scandi-2e4 49.5% ±1.7 Official leaderboard
246 Ministral 3 8B Instruct 49.4% ±1.2 val Dec 2025 Official leaderboard
247 OpenLLM-France/Luciole-23B-Instruct-1.1 49.3% ±1 val Official leaderboard
248 Qwen/Qwen3-4B#no-thinking 49.3% ±1 Official leaderboard
249 Qwen3-4B 49.3% ±1 Apr 2025 Official leaderboard
250 utter-project/EuroLLM-9B-Instruct-2512 49.3% ±0.6 val Official leaderboard
251 LumiOpen/Viking-13B 49.3% ±0.8 Official leaderboard
252 Unbabel/Tower-Plus-2B 49.2% ±1.2 val Official leaderboard
253 upstage/SOLAR-10.7B-v1.0 49.1% ±1.5 Official leaderboard
254 gpt-5-nano-2025-08-07#minimal 49.1% ±1.1 zero-shot, val Official leaderboard
255 KennethEnevoldsen/munin_mistral-7b 49.1% ±1.2 val Official leaderboard
256 Falcon3-10B-Instruct 49.1% ±0.9 Dec 2024 Official leaderboard
257 deepseek-r1-0528-qwen3-8b 49% ±1.6 May 2025 Official leaderboard
258 Gemma 2 27B 49% ±1.2 Jun 2024 Official leaderboard
259 gpt-5-mini-2025-08-07#minimal 49% ±1.1 zero-shot, val Official leaderboard
260 NbAiLab/borealis2-e2b-preview 48.9% ±1 val Official leaderboard
261 timpal0l/BeagleCatMunin2 48.8% ±1.4 val Official leaderboard
262 Ministral 3 3B Reasoning 48.7% ±1.1 Dec 2025 Official leaderboard
263 amalia-llm/AMALIA-9B-0626-DPO 48.7% ±1.2 zero-shot, val Official leaderboard
264 danish-foundation-models/munin-apertus-8b 48.7% ±0.7 Official leaderboard
265 synquid/gemma-3-4b-pt-laerebogen-with-follow-ups-lora 48.6% ±0.9 Official leaderboard
266 Tiny Aya Water 48.6% ±1 val Feb 2026 Official leaderboard
267 allenai/Olmo-3-7B-Think 48.5% ±0.6 Official leaderboard
268 meta-llama/Llama-2-13b-chat-hf 48.4% ±1.1 Official leaderboard
269 NbAiLab/borealis-4b-instruct-preview 48.2% ±0.4 Official leaderboard
270 NbAiLab/borealis-open-4b 48.2% ±0.9 val Official leaderboard
271 Granite-3.0-8B-Base 48.1% ±0.9 Oct 2024 Official leaderboard
272 CohereLabs/tiny-aya-base 48.1% ±1.5 val Official leaderboard
273 vesteinn/DanskBERT 48.1% ±1.2 Official leaderboard
274 amalia-llm/AMALIA-9B-0626-SFT 48% ±1.3 val Official leaderboard
275 Llama 3.1-8B 48% ±1.3 Jul 2024 Official leaderboard
276 mistralai/Ministral-3-3B-Base-2512 48% ±1 Official leaderboard
277 nvidia-nemotron-nano-9b-v2 48% ±0.6 Aug 2025 Official leaderboard
278 granite-4.0-micro 48% ±1 Oct 2025 Official leaderboard
279 Tiny Aya Global 47.9% ±0.9 val Feb 2026 Official leaderboard
280 Falcon3-7B-Instruct 47.8% ±0.8 Dec 2024 Official leaderboard
281 swiss-ai/Apertus-v1.1-4B 47.7% ±1.6 val Official leaderboard
282 allenai/OLMo-2-1124-13B-Instruct 47.7% ±0.9 Official leaderboard
283 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16#no-thinking 47.6% ±1.5 Official leaderboard
284 mistralai/Ministral-3-8B-Base-2512 47.6% ±1.6 val Official leaderboard
285 neph1/bellman-7b-mistral-instruct-v0.2 47.6% ±0.7 Official leaderboard
286 allenai/Llama-3.1-Tulu-3-8B-SFT 47.5% ±0.9 Official leaderboard
287 sentence-transformers/use-cmlm-multilingual 47.5% ±1.1 Official leaderboard
288 tiiuae/Falcon3-10B-Base 47.5% ±0.7 Official leaderboard
289 qwen3-4b-instruct-2507 47.4% ±0.7 Aug 2025 Official leaderboard
290 Mistral 7B v0.3 47.4% ±0.9 Sep 2023 Official leaderboard
291 mlabonne/NeuralBeagle14-7B 47.4% ±1.6 val Official leaderboard
292 vesteinn/FoBERT 47.4% ±1 Official leaderboard
293 Gemma 7B 47.3% ±1.2 Feb 2024 Official leaderboard
294 openeurollm/prelude-long 47.3% ±1.1 val Official leaderboard
295 occiglot/occiglot-7b-eu5 47.2% ±1 Official leaderboard
296 bineric/lynx-instruct-30b#no-thinking 47.2% ±1 Official leaderboard
297 Gemma 2 2B IT 47.1% ±0.9 Jul 2024 Official leaderboard
298 domyn/Domyn-Small-v1.0 47.1% ±1.5 val Official leaderboard
299 DGurgurov/xlm-r_dan-latn 46.9% ±1.3 Official leaderboard
300 norallm/normistral-11b-long 46.9% ±0.9 Official leaderboard
301 bineric/lynx-instruct-30b#thinking 46.8% ±1.1 Official leaderboard
302 openeurollm/oellm-9b-128k-theta32m-v3 46.8% ±1.4 val Official leaderboard
303 openeurollm/prelude-base 46.7% ±1.9 val Official leaderboard
304 Yi-1.5-34B 46.6% ±1.4 May 2024 Official leaderboard
305 mistralai/Ministral-8B-Instruct-2410 46.3% ±1 Official leaderboard
306 mii-llm/nesso-4B 46.2% ±0.5 Official leaderboard
307 swiss-ai/Apertus-v1.1-4B-Instruct 46.2% ±0.9 val Official leaderboard
308 vesteinn/ScandiBERT-no-faroese 46.2% ±1.9 Official leaderboard
309 nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16 46.1% ±0.8 Official leaderboard
310 DeepSeek-R1-Distill-Qwen-14B 45.9% ±1 Jan 2025 Official leaderboard
311 Apertus 1.5 8B 45.8% ±1 zero-shot, val Jul 2026 Official leaderboard
312 Qwen3-1.7B 45.7% ±0.7 Apr 2025 Official leaderboard
313 allenai/OLMo-2-1124-7B-Instruct 45.6% ±0.9 Official leaderboard
314 Qwen/Qwen3-1.7B#no-thinking 45.5% ±0.7 Official leaderboard
315 Mistral 7B v0.1 45.2% ±1.3 Sep 2023 Official leaderboard
316 alexandrainst/Meta-Llama-3.1-8B-dialog-k2-da 45.1% ±1.2 zero-shot, val Official leaderboard
317 TrustLLMeu/baseline-7-8b_1t-tokens_llama 45.1% ±1.2 Official leaderboard
318 HPLT/NorOLMo-13B 45% ±1.6 val Official leaderboard
319 Mistral-7B-Instruct-v0.2 44.9% ±1.2 Dec 2023 Official leaderboard
320 Mistral-7B-Instruct-v0.1 44.7% ±0.9 Sep 2023 Official leaderboard
321 LiquidAI/LFM2.5-8B-A1B-Base 44.6% ±1.2 val Official leaderboard
322 tiiuae/Falcon3-7B-Base 44.6% ±0.9 Official leaderboard
323 birgermoell/Munin-NeuralBeagle-NorskGPT 44.5% ±1.4 val Official leaderboard
324 ai21labs/Jamba-v0.1 44.5% ±1.4 Official leaderboard
325 RuterNorway/Llama-2-13b-chat-norwegian 44.5% ±1.3 Official leaderboard
326 norallm/normistral-11b-thinking 44.4% ±0.7 Official leaderboard
327 baidu/ERNIE-4.5-21B-A3B-PT 44.3% ±0.5 Official leaderboard
328 allenai/Olmo-3-1025-7B 44.3% ±0.8 Official leaderboard
329 mistralai/Voxtral-Mini-3B-2507 44.2% ±1.1 Official leaderboard
330 swiss-ai/Apertus-8B-2509 44.1% ±1.3 Official leaderboard
331 setu4993/LaBSE 44.1% ±1 Official leaderboard
332 AI-Sweden-Models/ModernBERT-large 44% ±1.6 Official leaderboard
333 ltg/norbert3-large 44% ±2.1 Official leaderboard
334 deepseek-ai/DeepSeek-R1-Distill-Llama-8B 43.9% ±0.9 Official leaderboard
335 nvidia/NVIDIA-Nemotron-3-Nano-4B-FP8 43.8% ±2.3 val Official leaderboard
336 NbAiLab/nb-bert-ncc-male2female 43.8% ±0.9 Official leaderboard
337 NbAiLab/nb-roberta-base-ncc 43.7% ±0.7 Official leaderboard
338 Qwen1.5-14B 43.4% ±0.8 Feb 2024 Official leaderboard
339 norallm/normistral-11b-warm 43.1% ±0.9 Official leaderboard
340 deahmed/Qwen3.5-2B-da-task 43% ±0.7 val Official leaderboard
341 speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c 43% ±0.4 zero-shot Official leaderboard
342 NbAiLab/nb-bert-base 42.9% ±0.8 Official leaderboard
343 Qwen3.5-2B 42.7% ±1.2 val Feb 2026 Official leaderboard
344 mlabonne/AlphaMonarch-7B 42.6% ±1.7 val Official leaderboard
345 speakleash/Bielik-4.5B-v3.0-Instruct 42.6% ±1.3 Official leaderboard
346 occiglot/occiglot-7b-eu5-instruct 42.5% ±0.8 Official leaderboard
347 AFM-4.5B 42.5% ±0.8 Jul 2025 Official leaderboard
348 ltg/norbert3-base 42.3% ±0.6 Official leaderboard
349 meta-llama/Llama-3.2-3B-Instruct 42.2% ±0.5 Official leaderboard
350 NbAiLab/borealis-1b-instruct-preview 42.1% ±1 val Official leaderboard
351 syvai/claes-v1 41.9% ±0.4 zero-shot Official leaderboard
352 Gemma 3 1B 41.9% ±0.9 Mar 2025 Official leaderboard
353 NbAiLab/borealis-open-1b 41.9% ±1.2 val Official leaderboard
354 NbAiLab/notram-bert-norwegian-cased-080321 41.6% ±1 Official leaderboard
355 ibm-granite/granite-34b-code-instruct 41.5% ±0.7 Official leaderboard
356 Qwen/Qwen2.5-1.5B-Instruct 41.4% ±0.7 Official leaderboard
357 arcee-ai/AFM-4.5B-Base 41.3% ±0.7 Official leaderboard
358 ibm-granite/granite-4.0-h-tiny 41.3% ±1.3 Official leaderboard
359 Qwen3.5-0.8B 41% ±0.9 val Feb 2026 Official leaderboard
360 jhu-clsp/mmBERT-base 40.8% ±1.1 Official leaderboard
361 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f 40.8% ±0.8 Official leaderboard
362 Qwen/Qwen2.5-Math-7B 40.7% ±0.9 Official leaderboard
363 AI-Sweden-Models/bert-large-nordic-pile-1M-steps 40.7% ±1 Official leaderboard
364 CohereForAI/aya-23-8B 40.6% ±1 Official leaderboard
365 KennethEnevoldsen/dfm-sentence-encoder-small 40.5% ±0.6 Official leaderboard
366 ibm-granite/granite-8b-code-instruct-4k 40.4% ±1 Official leaderboard
367 OpenLLM-France/Luciole-8B-Instruct-1.1 40.4% ±2.2 val Official leaderboard
368 Phi-4-reasoning 40.4% ±1.9 Apr 2025 Official leaderboard
369 NbAiLab/borealis-1b 40% ±1.9 val Official leaderboard
370 common-pile/comma-v0.1-2t 39.7% ±0.7 Official leaderboard
371 HuggingFaceTB/SmolLM3-3B-Base 39.6% ±1.6 Official leaderboard
372 swiss-ai/Apertus-v1.1-1.5B 39.5% ±1.4 val Official leaderboard
373 FacebookAI/xlm-roberta-large 39.5% ±5.2 Official leaderboard
374 speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551 39.4% ±0.5 zero-shot Official leaderboard
375 Gemma 2 9B 39.3% ±1.5 Jun 2024 Official leaderboard
376 Apertus 8B Instruct 39.2% ±4.4 Sep 2025 Official leaderboard
377 utter-project/EuroVLM-1.7B-Preview 39.2% ±1.6 Official leaderboard
378 ibm-granite/granite-34b-code-base 39.1% ±1 Official leaderboard
379 meta-llama/Llama-2-7b-chat-hf 39% ±0.8 Official leaderboard
380 Gemma 2B 38.9% ±0.5 Feb 2024 Official leaderboard
381 infly/OpenCoder-8B-Instruct 38.8% ±0.8 Official leaderboard
382 NbAiLab/nb-bert-large 38.6% ±3.7 Official leaderboard
383 allenai/Olmo-3-7B-Instruct 38.5% ±1.3 Official leaderboard
384 danish-foundation-models/gemma-3-1b-cpt-dynaword-full-v1 38.5% ±0.8 Official leaderboard
385 BSC-LT/salamandra-7b-instruct 38.5% ±1.3 Official leaderboard
386 speakleash/Bielik-1.5B-v3.0-Instruct 38.4% ±1 Official leaderboard
387 ltg/norbert3-small 38.4% ±1.2 Official leaderboard
388 ibm-granite/granite-8b-code-base-4k 38.3% ±1.5 Official leaderboard
389 ibm-granite/granite-4.0-h-micro 38.2% ±1.2 Official leaderboard
390 swiss-ai/Apertus-v1.1-1.5B-Instruct 38.2% ±1.5 val Official leaderboard
391 allenai/Olmo-3-7B-Instruct-DPO 37.9% ±1.3 Official leaderboard
392 ltg/norbert2 37.9% ±1.7 Official leaderboard
393 danish-foundation-models/munin-7b-alpha 37.2% ±0.8 Official leaderboard
394 KBLab/megatron-bert-base-swedish-cased-600k 37.2% ±1.1 Official leaderboard
395 ibm-granite/granite-20b-code-base 37.1% ±0.6 Official leaderboard
396 norallm/normistral-7b-scratch 36.9% ±2.3 val Official leaderboard
397 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660 36.8% ±0.7 Official leaderboard
398 DDSC/roberta-base-danish 36.8% ±3.5 Official leaderboard
399 Gemma 3 1B PT 36.7% ±1.3 Mar 2025 Official leaderboard
400 google/gemma-2-2b 36.6% ±0.9 Official leaderboard
401 danish-foundation-models/munin-gemma4-e4b 36.5% ±0.9 Official leaderboard
402 OmniJev/OneJev1.6-0.8B 36.5% ±0.7 zero-shot Official leaderboard
403 ibm-granite/granite-3.0-3b-a800m-instruct 36.2% ±1.3 Official leaderboard
404 Hunyuan-4B-Instruct 36.1% ±0.7 Jul 2025 Official leaderboard
405 common-pile/comma-v0.1-1t 36% ±1 Official leaderboard
406 syvai/claes-preview 35.7% ±0.5 zero-shot Official leaderboard
407 LiquidAI/LFM2-2.6B 35.5% ±0.6 Official leaderboard
408 google/rembert 35.5% ±1.2 Official leaderboard
409 KBLab/bert-base-swedish-cased-new 35.5% ±1.2 Official leaderboard
410 Maltehb/danish-bert-botxo 35.4% ±1.6 Official leaderboard
411 danish-foundation-models/munin-7b-core-pt-3 35.4% ±1.2 Official leaderboard
412 HPLT/gpt-13b-nordic-prerelease 35% ±2.1 Official leaderboard
413 ibm-granite/granite-3b-code-base-2k 35% ±1 Official leaderboard
414 HuggingFaceTB/SmolLM3-3B 34.9% ±2 Official leaderboard
415 KBLab/megatron-bert-large-swedish-cased-165k 34.9% ±1.1 Official leaderboard
416 allenai/Olmo-3-7B-Instruct-SFT 34.5% ±1.5 Official leaderboard
417 yulan-team/YuLan-Mini-Instruct 34.3% ±0.9 Official leaderboard
418 allenai/OLMo-2-1124-7B 34.3% ±1.2 Official leaderboard
419 danish-foundation-models/munin-7b-core-it 34.2% ±1.3 Official leaderboard
420 tencent/Hunyuan-7B-Instruct 33.7% ±0.8 Official leaderboard
421 fastino/GLiNER2.5-Decide 33.2% ±0.5 zero-shot Official leaderboard
422 CohereLabs/aya-23-8B 33.1% ±1.9 val Official leaderboard
423 microsoft/mdeberta-v3-base 32.9% ±2.8 Official leaderboard
424 ltg/norbert3-xs 32.8% ±0.7 Official leaderboard
425 vesteinn/ScandiBERT 32.7% ±6.1 Official leaderboard
426 mideind/IceBERT-xlmr-ic3 32.5% ±2.5 Official leaderboard
427 MaLA-LM/emma-500-llama2-7b 32.2% ±3.4 Official leaderboard
428 KBLab/megatron-bert-base-swedish-cased-125k 31.9% ±1 Official leaderboard
429 TildeAI/TildeOpen-30b#slow-tokenizer 31.7% ±1.1 Official leaderboard
430 HuggingFaceTB/SmolLM3-3B#no-thinking 31.6% ±1.5 Official leaderboard
431 KBLab/bert-base-swedish-cased 31.6% ±0.8 Official leaderboard
432 jhu-clsp/mmBERT-small 31.5% ±1.7 Official leaderboard
433 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6 31.5% ±1.3 Official leaderboard
434 google-bert/bert-base-multilingual-cased 31% ±2.7 val Official leaderboard
435 danish-foundation-models/munin-7b-core-pt-2 30.8% ±1.8 Official leaderboard
436 Falcon3-3B-Instruct 30.8% ±0.7 Dec 2024 Official leaderboard
437 utter-project/EuroLLM-1.7B-Instruct 30.6% ±2 val Official leaderboard
438 danish-foundation-models/encoder-small-v1 30.1% ±1.5 Official leaderboard
439 google-bert/bert-base-multilingual-uncased 29.6% ±1.7 val Official leaderboard
440 TildeAI/TildeOpen-30b 29.6% ±1.9 val Official leaderboard
441 allenai/Olmo-3-1125-32B 29.4% ±0.6 Official leaderboard
442 KBLab/roberta-base-swedish-cased 29.3% ±1 Official leaderboard
443 nvidia/NVIDIA-Nemotron-Nano-9B-v2#no-thinking 28.8% ±1.3 Official leaderboard
444 norallm/norbloom-7b-scratch 28.7% ±1.9 val Official leaderboard
445 tiiuae/Falcon3-3B-Base 28.5% ±1.7 Official leaderboard
446 TurkuNLP/finnish-modernbert-tiny 28.2% ±0.5 Official leaderboard
447 PleIAs/Pleias-1.2b-Preview 28.1% ±1.5 Official leaderboard
448 Llama 3.2 1B 28% ±1.2 Sep 2024 Official leaderboard
449 Maltehb/aelaectra-danish-electra-small-uncased 27.8% ±1.2 Official leaderboard
450 distilbert/distilbert-base-multilingual-cased 27.8% ±0.9 Official leaderboard
451 Geotrend/distilbert-base-25lang-cased 27.7% ±1 Official leaderboard
452 google/gemma-7b-it 27.6% ±2.6 Official leaderboard
453 Qwen3-0.6B 27.4% ±0.9 Apr 2025 Official leaderboard
454 KBLab/sentence-bert-swedish-cased 27.4% ±1.5 Official leaderboard
455 meta-llama/Llama-2-13b-hf 27.4% ±2.3 Official leaderboard
456 NorwAI/NorwAI-Mixtral-8x7B-one-third-epoch 27.3% ±1.7 val Official leaderboard
457 BSC-LT/ALIA-40b 26.9% ±2.9 Official leaderboard
458 ibm-granite/granite-20b-code-instruct 26.9% ±0.6 Official leaderboard
459 allenai/OLMo-2-0325-32B 26.6% ±2 Official leaderboard
460 TurkuNLP/finnish-modernbert-large 26.4% ±1.4 Official leaderboard
461 TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49 25.9% ±1.3 Official leaderboard
462 LFM2.5-1.2B-Thinking 25.8% ±0.7 Jan 2026 Official leaderboard
463 convaiinnovations/laya-multilingual 25.7% ±0.6 zero-shot Official leaderboard
464 TurkuNLP/finnish-modernbert-large-seq-len-1024-117300-annealed 25.3% ±1.6 Official leaderboard
465 Qwen3.5-9B-Base 24.3% ±2.9 val Feb 2026 Official leaderboard
466 swiss-ai/Apertus-v1.1-0.5B-Instruct 24.1% ±1.9 val Official leaderboard
467 swiss-ai/Apertus-v1.1-0.5B 24.1% ±2.9 val Official leaderboard
468 Maltehb/aelaectra-danish-electra-small-cased 23.9% ±2.9 Official leaderboard
469 TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894 23.8% ±1.9 Official leaderboard
470 tencent/Hunyuan-1.8B-Instruct 23.5% ±0.8 Official leaderboard
471 EuropeanParliament/EUBERT 23.4% ±1.1 Official leaderboard
472 LFM2.5-1.2B-Instruct 23.2% ±1.4 Jan 2026 Official leaderboard
473 neurocode/IsRoBERTa 23% ±0.7 Official leaderboard
474 KBLab/bert-base-swedish-cased-alpha 22.6% ±0.8 Official leaderboard
475 OpenLLM-France/Luciole-1B-Instruct-1.1 22.5% ±1.4 val Official leaderboard
476 OpenLLM-France/Luciole-23B-Base 22.5% ±1.8 val Official leaderboard
477 EuroBERT/EuroBERT-610m 22.5% ±1.4 Official leaderboard
478 EuroBERT/EuroBERT-210m 22.3% ±1.1 Official leaderboard
479 google/gemma-2b-it 22.3% ±1.6 Official leaderboard
480 TurkuNLP/finnish-modernbert-base 22.2% ±1 Official leaderboard
481 timpal0l/Mistral-7B-v0.1-flashback-v2 22.1% ±2.8 Official leaderboard
482 vesteinn/XLMR-ENIS 21.2% ±2.8 Official leaderboard
483 arcee-ai/Trinity-Nano-Preview 21.1% ±2.1 Official leaderboard
484 marksverdhei/modern-norbert3-large 20.7% ±2.5 Official leaderboard
485 answerdotai/ModernBERT-base 20.6% ±0.9 Official leaderboard
486 mideind/IceBERT-large 20.5% ±1.1 Official leaderboard
487 danish-foundation-models/munin-7b-core-pt 20.3% ±2.1 Official leaderboard
488 jaredpalmer/kev-0.5b 20.1% ±0.6 zero-shot Official leaderboard
489 mideind/IceBERT-igc 20.1% ±1.2 Official leaderboard
490 DeepSeek-R1-Distill-Qwen-7B 20% ±1.7 Jan 2025 Official leaderboard
491 Qwen/Qwen3-0.6B#no-thinking 19.8% ±2 Official leaderboard
492 Phi-2 19.4% ±1.6 Dec 2023 Official leaderboard
493 EuroBERT/EuroBERT-2.1B 19.4% ±1.5 Official leaderboard
494 Phi-4-mini-reasoning 19.2% ±1.6 Apr 2025 Official leaderboard
495 answerdotai/ModernBERT-large 19.1% ±1.3 Official leaderboard
496 arcee-ai/Trinity-Mini-Base 19% ±2 Official leaderboard
497 speakleash/Bielik-11B-v3-Base-20250730 18.8% ±2.4 Official leaderboard
498 danish-foundation-models/munin-7b-open-pt 18.6% ±3.5 Official leaderboard
499 mideind/IceBERT 18.2% ±0.4 Official leaderboard
500 mideind/IceBERT-ic3 17.5% ±1.5 Official leaderboard
501 DeepSeek-R1-Distill-Qwen-1.5B 17.4% ±1.1 Jan 2025 Official leaderboard
502 AI-Sweden-Models/gpt-sw3-126m-instruct 17.4% ±1.7 Official leaderboard
503 LiquidAI/LFM2-350M 17.2% ±1.3 Official leaderboard
504 facebook/MobileLLM-R1-950M-base 17.2% ±1.1 Official leaderboard
505 allenai/OLMo-2-1124-13B 17.1% ±1.8 Official leaderboard
506 AI-Sweden-Models/gpt-sw3-1.3b-instruct 16.9% ±1.9 val Official leaderboard
507 LFM2-1.2B 16.9% ±1.5 Jul 2025 Official leaderboard
508 PleIAs/Pleias-3b-Preview 16.6% ±2.1 Official leaderboard
509 utter-project/EuroMoE-2.6B-A0.6B-Preview 16.5% ±2.9 Official leaderboard
510 Gemma 3 270M 16.2% ±1.6 Aug 2025 Official leaderboard
511 utter-project/EuroLLM-22B-2512 16.1% ±1.2 Official leaderboard
512 NorwAI/NorwAI-Mixtral-8x7B 16% ±2.5 val Official leaderboard
513 fresh-electra-small 16% ±2.7 Official leaderboard
514 fresh-xlm-roberta-base 16% ±1.3 Official leaderboard
515 LiquidAI/LFM2.5-1.2B-Base 15.8% ±2.4 Official leaderboard
516 DDSC/roberta-base-scandinavian 15.8% ±3.2 Official leaderboard
517 OpenLLM-France/Luciole-8B-Base 15.7% ±2.2 val Official leaderboard
518 SmolLM2 1.7B Instruct 15.7% ±1.4 Oct 2024 Official leaderboard
519 DiscoResearch/DiscoLM-70b 15.6% ±2.7 Official leaderboard
520 infly/OpenCoder-8B-Base 15.4% ±4.5 Official leaderboard
521 marksverdhei/modern-norbert3-xs 15.3% ±1 Official leaderboard
522 ibm-granite/granite-3b-code-instruct-2k 15.2% ±1.2 Official leaderboard
523 marksverdhei/modern-norbert3-base 15% ±1.2 Official leaderboard
524 openGPT-X/Teuken-7B-instruct-commercial-v0.4 14.8% ±2.1 Official leaderboard
525 NbAiLab/roberta_jan_128_ncc 14.3% ±1.3 Official leaderboard
526 convaiinnovations/laya 14% ±0.5 zero-shot Official leaderboard
527 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview 13.1% ±2.7 Official leaderboard
528 RJuro/kanelsnegl-v0.1 13% ±2.1 Official leaderboard
529 infly/OpenCoder-1.5B-Instruct 12.9% ±2 Official leaderboard
530 LiquidAI/LFM2-700M 12.6% ±1.3 Official leaderboard
531 HPLT/hplt2c_swe_checkpoints 12.4% ±0.9 Official leaderboard
532 PleIAs/Pleias-350m-Preview 10.5% ±1.5 Official leaderboard
533 arcee-ai/Trinity-Nano-Base 10.5% ±2.5 Official leaderboard
534 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base 9.75% ±1.6 Official leaderboard
535 marksverdhei/modern-norbert3-small 9.61% ±1.3 Official leaderboard
536 ibm-granite/granite-3.0-3b-a800m-base 8.98% ±1.1 Official leaderboard
537 SmolLM2-1.7B 8.53% ±1.4 Oct 2024 Official leaderboard
538 HuggingFaceTB/SmolLM2-135M-Instruct 8.14% ±1.5 Official leaderboard
539 HPLT/hplt2c_isl_checkpoints 7.71% ±2.2 Official leaderboard
540 mideind/IceBERT-mC4-is 7.51% ±3.1 Official leaderboard
541 Qwen3.5-4B-Base 6.93% ±1.4 val Feb 2026 Official leaderboard
542 PleIAs/Monad 6.80% ±2 Official leaderboard
543 LFM2.5-8B-A1B 6.70% ±2.1 val May 2026 Official leaderboard
544 Phi-4-Reasoning-plus 6.42% ±3.3 Apr 2025 Official leaderboard
545 meta-llama/Llama-3.2-1B-Instruct 6.38% ±2.6 Official leaderboard
546 Qwen3.5-2B-Base 6.16% ±2.6 val Feb 2026 Official leaderboard
547 HPLT/hplt2c_nob_checkpoints 5.74% ±2.6 Official leaderboard
548 NbAiLab/borealis-270m 5.49% ±2.2 val Official leaderboard
549 BSC-LT/salamandra-7b 5.14% ±1.4 Official leaderboard
550 SmolLM2-360M 4.99% ±1.3 Oct 2024 Official leaderboard
551 NbAiLab/borealis-open-270m 4.84% ±1.9 val Official leaderboard
552 Qwen3.5-0.8B-Base 4.45% ±2 val Feb 2026 Official leaderboard
553 openai-community/gpt2 3.67% ±1 Official leaderboard
554 AI-Sweden-Models/gpt-sw3-356m-instruct 3.64% ±0.7 Official leaderboard
555 SmolLM2-135M 3.55% ±1.3 Oct 2024 Official leaderboard
556 AI-Sweden-Models/gpt-sw3-6.7b-v2-instruct 3.44% ±1.4 val Official leaderboard
557 baidu/ERNIE-4.5-0.3B-PT 3.29% ±0.9 Official leaderboard
558 infly/OpenCoder-1.5B-Base 2.78% ±0.9 Official leaderboard
559 AI-Sweden-Models/gpt-sw3-40b 2.69% ±1.2 val Official leaderboard
560 openbmb/MiniCPM5-1B 2.52% ±2.5 val Official leaderboard
561 openai-community/gpt2-medium 2.29% ±1.4 Official leaderboard
562 NbAiLab/borealis-270m-instruct-preview 2.13% ±1.1 val Official leaderboard
563 kyutai/helium-1-2b 1.77% ±0.6 Official leaderboard
564 PleIAs/Baguettotron@refs/pr/6#no-thinking 1.74% ±0.9 Official leaderboard
565 Muse-research/Muse-3B 1.62% ±2 val Official leaderboard
566 openai-community/gpt2-large 1.44% ±1.5 Official leaderboard
567 OpenLLM-France/Luciole-1B-Base 1.16% ±0.8 val Official leaderboard
568 HPLT/hplt2c_dan_checkpoints 0.96% ±1.1 Official leaderboard
569 HuggingFaceTB/SmolLM2-360M-Instruct 0.77% ±0.6 Official leaderboard
570 AI-Sweden-Models/gpt-sw3-6.7b-v2-translator 0.50% ±0.5 val Official leaderboard
571 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct 0.27% ±0.5 Official leaderboard
572 facebook/MobileLLM-R1-360M 0.25% ±1.5 Official leaderboard
573 ltg/norbert 0.15% ±1.5 Official leaderboard
574 facebook/MobileLLM-R1-140M 0.12% ±0.4 Official leaderboard
575 AI-Sweden-Models/gpt-sw3-1.3b 0% val Official leaderboard
576 AI-Sweden-Models/gpt-sw3-20b 0% Official leaderboard
577 AI-Sweden-Models/gpt-sw3-356m 0% Official leaderboard
578 AI-Sweden-Models/gpt-sw3-6.7b 0% val Official leaderboard
579 AI-Sweden-Models/gpt-sw3-6.7b-v2 0% val Official leaderboard
580 facebook/MobileLLM-R1-140M-base 0% Official leaderboard
581 facebook/MobileLLM-R1-950M 0% Official leaderboard
582 NorwAI/NorwAI-Magistral-24B-reasoning 0% val Official leaderboard
583 openGPT-X/Teuken-7B-base-v0.6 0% val Official leaderboard
584 Qwen/Qwen2.5-Math-1.5B 0% Official leaderboard
585 utter-project/EuroLLM-1.7B 0% val Official leaderboard
586 PleIAs/Baguettotron@refs/pr/6#thinking -0.02% ±0 Official leaderboard
587 facebook/MobileLLM-R1-360M-base -0.35% ±0.2 Official leaderboard
588 AI-Sweden-Models/gpt-sw3-126m -0.37% ±0.2 Official leaderboard
589 danish-foundation-models/munin-7b-open-it -0.47% ±1.3 Official leaderboard
590 HPLT/hplt2c_nno_checkpoints -0.72% ±0.5 Official leaderboard
591 inclusionAI/Ling-mini-base-2.0 -0.89% ±0.9 val Official leaderboard
592 Shekswess/trlm-135m -0.98% ±0.7 Official leaderboard
593 PleIAs/Baguettotron@refs/pr/6 -1.12% ±0.8 Official leaderboard
594 inclusionAI/Ling-mini-2.0 -1.15% ±2.4 val Official leaderboard
595 norallm/normistral-11b-translate -1.60% ±1 val Official leaderboard
596 AIDC-AI/Marco-Mini-Global-Base -2.03% ±1.2 val Official leaderboard
#ModelScore ±SettingDateSource
1 GPT-6.2 Astra 67% ±1.1 zero-shot, val Nov 2026 Official leaderboard
2 GLM-5.3-Flash 66.3% ±0.8 val Aug 2026 Official leaderboard
3 GPT-6 Astra 65.3% ±1.3 zero-shot, val Sep 2026 Official leaderboard
4 Gemini 3.7 Flash 64.2% ±0.9 zero-shot, val Aug 2026 Official leaderboard
5 Gemini 2.5 Flash Sep 2025 63.8% ±0.9 zero-shot, val Sep 2025 Official leaderboard
6 Gemma 4 31B IT 63.5% ±1 val Apr 2026 Official leaderboard
7 GPT-6.2 Sol 62.8% ±1.3 zero-shot, val Nov 2026 Official leaderboard
8 GPT-6.1 Sol 62.6% ±1 zero-shot, val Sep 2026 Official leaderboard
9 Mistral Medium 3.5 62.2% ±1.1 val Apr 2026 Official leaderboard
10 o3-mini 62.2% ±0.9 zero-shot, val Jan 2025 Official leaderboard
11 Gemini 3.5 Flash 61.9% ±1 zero-shot, val May 2026 Official leaderboard
12 Qwen3.5 397B-A17B 60.9% ±0.6 Feb 2026 Official leaderboard
13 Claude Opus 4.8 60.6% ±0.8 zero-shot, val May 2026 Official leaderboard
14 Qwen 3.8 27B 60.6% ±1.1 val Aug 2026 Official leaderboard
15 GPT-6 Luna 60.2% ±1.3 zero-shot, val Sep 2026 Official leaderboard
16 Gemini 3.6 Flash 60.2% ±0.9 zero-shot, val Jul 2026 Official leaderboard
17 Gemini 3 Pro 60.2% ±0.7 zero-shot, val Nov 2025 Official leaderboard
18 Olmo 3.1 32B Think 60% ±1.3 val Dec 2025 Official leaderboard
19 Gemma 4 12B 60% ±0.9 val Jun 2026 Official leaderboard
20 GPT-5.6 Sol 59.3% ±1.3 zero-shot, val Jul 2026 Official leaderboard
21 GPT-5 58.8% ±1 zero-shot, val Aug 2025 Official leaderboard
22 Reka Flash 3.1 58.7% ±0.6 May 2025 Official leaderboard
23 Cogito v1 Preview 70B 58.6% ±0.7 Apr 2025 Official leaderboard
24 GPT-5.2 58.5% ±0.6 zero-shot, val Dec 2025 Official leaderboard
25 glm-4.7-flash 58.5% ±1.2 val Jan 2026 Official leaderboard
26 Llama 3.1-70B 58.4% ±0.9 Jul 2024 Official leaderboard
27 Qwen3.5-9B 58.2% ±1.9 val Feb 2026 Official leaderboard
28 Gemini 3.1 Flash-Lite 58.1% ±0.9 zero-shot, val Mar 2026 Official leaderboard
29 o3 58.1% ±1 zero-shot, val Apr 2025 Official leaderboard
30 Llama 4 Scout 58.1% ±0.4 Apr 2025 Official leaderboard
31 Reka Flash 3 58% ±0.7 Mar 2025 Official leaderboard
32 EXAONE 4.0 32B 58% ±1.4 val Jul 2025 Official leaderboard
33 Claude Sonnet 4.6 57.7% ±0.9 zero-shot, val Feb 2026 Official leaderboard
34 o4-mini 57.5% ±0.6 zero-shot, val Apr 2025 Official leaderboard
35 Gemma 3 27B 57.4% ±0.8 Mar 2025 Official leaderboard
36 GPT-5.5 57% ±1.4 zero-shot, val Apr 2026 Official leaderboard
37 Gemma 4 26B A4B 56.9% ±0.8 val Apr 2026 Official leaderboard
38 GPT-4.1 56.9% ±1 zero-shot, val Apr 2025 Official leaderboard
39 Nemotron-Cascade 2 30B-A3B 56.9% ±1.2 val Mar 2026 Official leaderboard
40 Llama 3.3 70B Instruct 56.8% ±0.7 zero-shot, val Dec 2024 Official leaderboard
41 Gemma 3 12B 56.7% ±0.7 Mar 2025 Official leaderboard
42 QwQ-32B 56.6% ±0.4 Mar 2025 Official leaderboard
43 Apertus 1.5 70B 56.4% ±1.3 val Jul 2026 Official leaderboard
44 OLMo 2 32B Instruct 56.2% ±0.9 Mar 2025 Official leaderboard
45 GLM-4.5-Air 56.2% ±0.6 thinking Jul 2025 Official leaderboard
46 Qwen3.5-35B-A3B-Base 56.1% ±1.2 val Feb 2026 Official leaderboard
47 Aya Expanse 8B 56% ±1.4 val Oct 2024 Official leaderboard
48 GPT-5 nano 56% ±0.9 zero-shot, val Aug 2025 Official leaderboard
49 Claude Haiku 4.5 55.9% ±0.8 zero-shot, val Oct 2025 Official leaderboard
50 Grok 4.20 55.9% ±0.8 thinking, zero-shot, val Feb 2026 Official leaderboard
51 Gemini 2.5 Pro Jun 2025 55.7% ±0.9 zero-shot, val Jun 2025 Official leaderboard
52 gpt-oss-120b 55.6% ±0.8 val Aug 2025 Official leaderboard
53 Qwen3-14B 55.5% ±0.8 Apr 2025 Official leaderboard
54 Ministral 3 3B Instruct 55.5% ±0.8 zero-shot, val Dec 2025 Official leaderboard
55 GPT-5.6 Terra 55.5% ±1.4 zero-shot, val Jul 2026 Official leaderboard
56 Apertus 70B 55.4% ±0.7 Sep 2025 Official leaderboard
57 Gemma 3n E4B 55.4% ±1 Jun 2025 Official leaderboard
58 Qwen3-235B-A22B 55.4% ±0.6 Apr 2025 Official leaderboard
59 Gemma 3n E2B 55.2% ±0.7 Jun 2025 Official leaderboard
60 GPT-6 Sol 55.2% ±1.4 zero-shot, val Sep 2026 Official leaderboard
61 Gemini 3.5 Flash-Lite 55.1% ±1.1 zero-shot, val Jul 2026 Official leaderboard
62 Qwen3-Next-80B-A3B 55.1% ±0.8 thinking Sep 2025 Official leaderboard
63 Grok-2 Dec 2024 55% ±1.5 zero-shot, val Dec 2024 Official leaderboard
64 Phi-4 55% ±0.7 Dec 2024 Official leaderboard
65 Grok 3 54.8% ±1 zero-shot, val Apr 2025 Official leaderboard
66 Voxtral Small 54.6% ±0.9 Jul 2025 Official leaderboard
67 Qwen3-8B 54.5% ±0.9 Apr 2025 Official leaderboard
68 gpt-oss-20b 54.5% ±0.8 high Aug 2025 Official leaderboard
69 Grok-3 mini 54.5% ±1 zero-shot, val Apr 2025 Official leaderboard
70 Gemma 3 27B PT 54.3% ±0.9 Mar 2025 Official leaderboard
71 Grok 4.1 Fast 54.1% ±1 zero-shot, val Nov 2025 Official leaderboard
72 Aya Expanse 32B 54.1% ±0.7 Oct 2024 Official leaderboard
73 GLM-4.1V-9B-Thinking 54% ±0.7 Jul 2025 Official leaderboard
74 Qwen3-Next-80B-A3B-Instruct 53.3% ±0.7 Sep 2025 Official leaderboard
75 Gemini 2.5 Flash-Lite Sep 2025 53.2% ±1.2 zero-shot, val Sep 2025 Official leaderboard
76 Gemma 3 12B PT 53% ±0.7 Mar 2025 Official leaderboard
77 Ministral 3 14B 52.8% ±1 val Dec 2025 Official leaderboard
78 DeepSeek-R1-Distill-Llama-70B 52.8% ±0.6 Jan 2025 Official leaderboard
79 Gemma 4 E4B 52.8% ±1.2 val Mar 2026 Official leaderboard
80 GPT-5 mini 52.7% ±0.9 zero-shot, val Aug 2025 Official leaderboard
81 Llama-3.1-Tulu-3-8B 52.6% ±1 Nov 2024 Official leaderboard
82 GPT-5.6 Luna 52.5% ±1.1 zero-shot, val Jul 2026 Official leaderboard
83 Olmo 3.1 32B Instruct 52.3% ±0.9 Dec 2025 Official leaderboard
84 Command R7B 52.1% ±0.5 Dec 2024 Official leaderboard
85 Qwen3-4B-Thinking-2507 52.1% ±0.8 Aug 2025 Official leaderboard
86 Llama 3.1-405B 51.9% ±0.7 Jul 2024 Official leaderboard
87 Granite 4.0 H Small 51.8% ±1 Oct 2025 Official leaderboard
88 Trinity Mini 51.1% ±0.6 Dec 2025 Official leaderboard
89 Qwen1.5-14B-Chat 51% ±0.7 Feb 2024 Official leaderboard
90 Granite-3.0-8B-Instruct 50.8% ±1.1 Oct 2024 Official leaderboard
91 Qwen3.5-4B 50.8% ±1.8 val Feb 2026 Official leaderboard
92 Gemma 4 E2B 49.5% ±1.2 val Mar 2026 Official leaderboard
93 DeepSeek-R1-Distill-Qwen-32B 49.5% ±0.8 Jan 2025 Official leaderboard
94 Ministral 3 8B Instruct 49.4% ±1.2 val Dec 2025 Official leaderboard
95 Qwen3-4B 49.3% ±1 Apr 2025 Official leaderboard
96 Falcon3-10B-Instruct 49.1% ±0.9 Dec 2024 Official leaderboard
97 deepseek-r1-0528-qwen3-8b 49% ±1.6 May 2025 Official leaderboard
98 Gemma 2 27B 49% ±1.2 Jun 2024 Official leaderboard
99 Ministral 3 3B Reasoning 48.7% ±1.1 Dec 2025 Official leaderboard
100 Tiny Aya Water 48.6% ±1 val Feb 2026 Official leaderboard
101 Granite-3.0-8B-Base 48.1% ±0.9 Oct 2024 Official leaderboard
102 Llama 3.1-8B 48% ±1.3 Jul 2024 Official leaderboard
103 nvidia-nemotron-nano-9b-v2 48% ±0.6 Aug 2025 Official leaderboard
104 granite-4.0-micro 48% ±1 Oct 2025 Official leaderboard
105 Tiny Aya Global 47.9% ±0.9 val Feb 2026 Official leaderboard
106 Falcon3-7B-Instruct 47.8% ±0.8 Dec 2024 Official leaderboard
107 qwen3-4b-instruct-2507 47.4% ±0.7 Aug 2025 Official leaderboard
108 Mistral 7B v0.3 47.4% ±0.9 Sep 2023 Official leaderboard
109 Gemma 7B 47.3% ±1.2 Feb 2024 Official leaderboard
110 Gemma 2 2B IT 47.1% ±0.9 Jul 2024 Official leaderboard
111 Yi-1.5-34B 46.6% ±1.4 May 2024 Official leaderboard
112 DeepSeek-R1-Distill-Qwen-14B 45.9% ±1 Jan 2025 Official leaderboard
113 Apertus 1.5 8B 45.8% ±1 zero-shot, val Jul 2026 Official leaderboard
114 Qwen3-1.7B 45.7% ±0.7 Apr 2025 Official leaderboard
115 Mistral 7B v0.1 45.2% ±1.3 Sep 2023 Official leaderboard
116 Mistral-7B-Instruct-v0.2 44.9% ±1.2 Dec 2023 Official leaderboard
117 Mistral-7B-Instruct-v0.1 44.7% ±0.9 Sep 2023 Official leaderboard
118 Qwen1.5-14B 43.4% ±0.8 Feb 2024 Official leaderboard
119 Qwen3.5-2B 42.7% ±1.2 val Feb 2026 Official leaderboard
120 AFM-4.5B 42.5% ±0.8 Jul 2025 Official leaderboard
121 Gemma 3 1B 41.9% ±0.9 Mar 2025 Official leaderboard
122 Qwen3.5-0.8B 41% ±0.9 val Feb 2026 Official leaderboard
123 Phi-4-reasoning 40.4% ±1.9 Apr 2025 Official leaderboard
124 Gemma 2 9B 39.3% ±1.5 Jun 2024 Official leaderboard
125 Apertus 8B Instruct 39.2% ±4.4 Sep 2025 Official leaderboard
126 Gemma 2B 38.9% ±0.5 Feb 2024 Official leaderboard
127 Gemma 3 1B PT 36.7% ±1.3 Mar 2025 Official leaderboard
128 Hunyuan-4B-Instruct 36.1% ±0.7 Jul 2025 Official leaderboard
129 Falcon3-3B-Instruct 30.8% ±0.7 Dec 2024 Official leaderboard
130 Llama 3.2 1B 28% ±1.2 Sep 2024 Official leaderboard
131 Qwen3-0.6B 27.4% ±0.9 Apr 2025 Official leaderboard
132 LFM2.5-1.2B-Thinking 25.8% ±0.7 Jan 2026 Official leaderboard
133 Qwen3.5-9B-Base 24.3% ±2.9 val Feb 2026 Official leaderboard
134 LFM2.5-1.2B-Instruct 23.2% ±1.4 Jan 2026 Official leaderboard
135 DeepSeek-R1-Distill-Qwen-7B 20% ±1.7 Jan 2025 Official leaderboard
136 Phi-2 19.4% ±1.6 Dec 2023 Official leaderboard
137 Phi-4-mini-reasoning 19.2% ±1.6 Apr 2025 Official leaderboard
138 DeepSeek-R1-Distill-Qwen-1.5B 17.4% ±1.1 Jan 2025 Official leaderboard
139 LFM2-1.2B 16.9% ±1.5 Jul 2025 Official leaderboard
140 Gemma 3 270M 16.2% ±1.6 Aug 2025 Official leaderboard
141 SmolLM2 1.7B Instruct 15.7% ±1.4 Oct 2024 Official leaderboard
142 SmolLM2-1.7B 8.53% ±1.4 Oct 2024 Official leaderboard
143 Qwen3.5-4B-Base 6.93% ±1.4 val Feb 2026 Official leaderboard
144 LFM2.5-8B-A1B 6.70% ±2.1 val May 2026 Official leaderboard
145 Phi-4-Reasoning-plus 6.42% ±3.3 Apr 2025 Official leaderboard
146 Qwen3.5-2B-Base 6.16% ±2.6 val Feb 2026 Official leaderboard
147 SmolLM2-360M 4.99% ±1.3 Oct 2024 Official leaderboard
148 Qwen3.5-0.8B-Base 4.45% ±2 val Feb 2026 Official leaderboard
149 SmolLM2-135M 3.55% ±1.3 Oct 2024 Official leaderboard

More effort adds a median 6 points

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
GLM-5.3-Flash
zero-shot, … val
66.3%
Mistral Medium 3.5
zero-shot, … val
62.2%
Qwen3.5 397B-A17B
zero-shot, …
60.9%
Apertus 1.5 70B
zero-shot, … val
56.4%
Aya Expanse 8B
val
56%
Ministral 3 3B Instruct
val
55.5%
gpt-oss-20b
low medium
54.5%
Apertus 1.5 8B
val
45.8%

Source: Official Angry Tweets leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
GPT-6.2 Astra zero-shot, val 67%
GLM-5.3-Flash val 66.3%
GPT-6 Astra zero-shot, val 65.3%
Gemini 3.7 Flash zero-shot, val 64.2%
Gemini 2.5 Flash Sep 2025 zero-shot, val 63.8%
Gemma 4 31B IT val 63.5%
GPT-6.2 Sol zero-shot, val 62.8%
GPT-6.1 Sol zero-shot, val 62.6%
Mistral Medium 3.5 val 62.2%
o3-mini zero-shot, val 62.2%
Gemini 3.5 Flash zero-shot, val 61.9%
Qwen3.5 397B-A17B 60.9%
Qwen3.5 397B-A17B zero-shot, val 60.7%
Claude Opus 4.8 zero-shot, val 60.6%
Qwen 3.8 27B val 60.6%
GPT-6 Luna zero-shot, val 60.2%
Gemini 3.6 Flash zero-shot, val 60.2%
Gemini 3 Pro zero-shot, val 60.2%
Olmo 3.1 32B Think val 60%
Gemma 4 12B val 60%
GPT-5.6 Sol zero-shot, val 59.3%
GPT-5 zero-shot, val 58.8%
Reka Flash 3.1 58.7%
GLM-5.3-Flash zero-shot, val 58.6%
Cogito v1 Preview 70B 58.6%
GPT-5.2 zero-shot, val 58.5%
glm-4.7-flash val 58.5%
Llama 3.1-70B 58.4%
Qwen3.5-9B val 58.2%
Gemini 3.1 Flash-Lite zero-shot, val 58.1%
o3 zero-shot, val 58.1%
Llama 4 Scout 58.1%
Reka Flash 3 58%
EXAONE 4.0 32B val 58%
Claude Sonnet 4.6 zero-shot, val 57.7%
o4-mini zero-shot, val 57.5%
Gemma 3 27B 57.4%
GPT-5.5 zero-shot, val 57%
Gemma 4 26B A4B val 56.9%
GPT-4.1 zero-shot, val 56.9%
Nemotron-Cascade 2 30B-A3B val 56.9%
Llama 3.3 70B Instruct zero-shot, val 56.8%
Gemma 3 12B 56.7%
QwQ-32B 56.6%
Apertus 1.5 70B val 56.4%
OLMo 2 32B Instruct 56.2%
GLM-4.5-Air thinking 56.2%
Qwen3.5-35B-A3B-Base val 56.1%
Aya Expanse 8B val 56%
GPT-5 nano zero-shot, val 56%
Claude Haiku 4.5 zero-shot, val 55.9%
Grok 4.20 thinking, zero-shot, val 55.9%
Gemini 2.5 Pro Jun 2025 zero-shot, val 55.7%
gpt-oss-120b val 55.6%
Gemini 2.5 Flash Sep 2025 zero-shot, val 55.5%
Qwen3-14B 55.5%
Ministral 3 3B Instruct zero-shot, val 55.5%
GPT-5.6 Terra zero-shot, val 55.5%
Apertus 70B 55.4%
Gemma 3n E4B 55.4%
Qwen3-235B-A22B 55.4%
Gemma 3n E2B 55.2%
GPT-6 Sol zero-shot, val 55.2%
Gemini 3.5 Flash-Lite zero-shot, val 55.1%
Qwen3-Next-80B-A3B thinking 55.1%
Grok-2 Dec 2024 zero-shot, val 55%
Phi-4 55%
Apertus 70B 54.8%
Grok 3 zero-shot, val 54.8%
Mistral Medium 3.5 zero-shot, val 54.8%
Voxtral Small 54.6%
Qwen3-8B 54.5%
gpt-oss-20b high 54.5%
Grok-3 mini zero-shot, val 54.5%
Gemma 3 27B PT 54.3%
Ministral 3 3B Instruct val 54.2%
Grok 4.1 Fast zero-shot, val 54.1%
Aya Expanse 32B 54.1%
GLM-4.1V-9B-Thinking 54%
gpt-oss-20b medium 53.7%
Qwen3-Next-80B-A3B-Instruct 53.3%
Gemini 2.5 Flash-Lite Sep 2025 zero-shot, val 53.2%
Gemma 3 12B PT 53%
Ministral 3 14B val 52.8%
DeepSeek-R1-Distill-Llama-70B 52.8%
Gemma 4 E4B val 52.8%
GPT-5 mini zero-shot, val 52.7%
Llama-3.1-Tulu-3-8B 52.6%
GPT-5.6 Luna zero-shot, val 52.5%
Olmo 3.1 32B Instruct 52.3%
Command R7B 52.1%
Qwen3-4B-Thinking-2507 52.1%
Llama 3.1-405B 51.9%
Granite 4.0 H Small 51.8%
Trinity Mini 51.1%
Qwen1.5-14B-Chat 51%
Granite-3.0-8B-Instruct 50.8%
Qwen3.5-4B val 50.8%
Aya Expanse 8B 49.7%
Gemma 4 E2B val 49.5%
DeepSeek-R1-Distill-Qwen-32B 49.5%
Ministral 3 8B Instruct val 49.4%
Apertus 1.5 70B zero-shot, val 49.4%
Qwen3-4B 49.3%
Falcon3-10B-Instruct 49.1%
deepseek-r1-0528-qwen3-8b 49%
Gemma 2 27B 49%
Ministral 3 3B Reasoning 48.7%
Tiny Aya Water val 48.6%
Granite-3.0-8B-Base 48.1%
Llama 3.1-8B 48%
gpt-oss-20b low 48%
nvidia-nemotron-nano-9b-v2 48%
granite-4.0-micro 48%
Tiny Aya Global val 47.9%
Falcon3-7B-Instruct 47.8%
qwen3-4b-instruct-2507 47.4%
Mistral 7B v0.3 47.4%
Gemma 7B 47.3%
Gemma 2 2B IT 47.1%
Gemma 3n E4B 47.1%
Yi-1.5-34B 46.6%
DeepSeek-R1-Distill-Qwen-14B 45.9%
Apertus 1.5 8B zero-shot, val 45.8%
Qwen3-1.7B 45.7%
Mistral 7B v0.1 45.2%
Mistral-7B-Instruct-v0.2 44.9%
Mistral-7B-Instruct-v0.1 44.7%
Apertus 1.5 8B val 44.1%
Qwen1.5-14B 43.4%
Qwen3.5-2B val 42.7%
AFM-4.5B 42.5%
Gemma 3 1B 41.9%
Qwen3.5-0.8B val 41%
Phi-4-reasoning 40.4%
Gemma 3 1B 40.3%
Gemma 4 31B IT val 40.1%
Gemma 2 9B 39.3%
Apertus 8B Instruct 39.2%
Gemma 2B 38.9%
Gemma 3 1B PT 36.7%
Hunyuan-4B-Instruct 36.1%
Gemma 4 26B A4B val 36%
Gemma 4 12B val 34.8%
Falcon3-3B-Instruct 30.8%
Llama 3.2 1B 28%
Qwen3-0.6B 27.4%
LFM2.5-1.2B-Thinking 25.8%
Qwen3.5-9B-Base val 24.3%
LFM2.5-1.2B-Instruct 23.2%
DeepSeek-R1-Distill-Qwen-7B 20%
Phi-2 19.4%
Phi-4-mini-reasoning 19.2%
DeepSeek-R1-Distill-Qwen-1.5B 17.4%
LFM2-1.2B 16.9%
Gemma 3 270M 16.2%
SmolLM2 1.7B Instruct 15.7%
Gemma 4 E2B val 12.4%
Gemma 4 E4B val 11.9%
SmolLM2-1.7B 8.53%
Qwen3.5-4B-Base val 6.93%
LFM2.5-8B-A1B val 6.70%
Phi-4-Reasoning-plus 6.42%
Gemma 3 270M 6.31%
Qwen3.5-2B-Base val 6.16%
SmolLM2-360M 4.99%
Qwen3.5-0.8B-Base val 4.45%
SmolLM2-135M 3.55%

More effort adds a median 6 points

Models the leaderboard ran at two or more effort levels. We don’t have enough cost data here to say what the extra effort costs.

Lowest effort to highest, light to dark
GLM-5.3-Flash
zero-shot, … val
66.3%
Mistral Medium 3.5
zero-shot, … val
62.2%
Qwen3.5 397B-A17B
zero-shot, …
60.9%
Apertus 1.5 70B
zero-shot, … val
56.4%
Aya Expanse 8B
val
56%
Ministral 3 3B Instruct
val
55.5%
gpt-oss-20b
low medium
54.5%
Apertus 1.5 8B
val
45.8%

Source: Official Angry Tweets leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
GPT-6.2 Astra zero-shot, val 67%
GLM-5.3-Flash val 66.3%
GPT-6 Astra zero-shot, val 65.3%
Gemini 3.7 Flash zero-shot, val 64.2%
Gemini 2.5 Flash Sep 2025 zero-shot, val 63.8%
Gemma 4 31B IT val 63.5%
GPT-6.2 Sol zero-shot, val 62.8%
GPT-6.1 Sol zero-shot, val 62.6%
Mistral Medium 3.5 val 62.2%
o3-mini zero-shot, val 62.2%
Gemini 3.5 Flash zero-shot, val 61.9%
Qwen3.5 397B-A17B 60.9%
Qwen3.5 397B-A17B zero-shot, val 60.7%
Claude Opus 4.8 zero-shot, val 60.6%
Qwen 3.8 27B val 60.6%
GPT-6 Luna zero-shot, val 60.2%
Gemini 3.6 Flash zero-shot, val 60.2%
Gemini 3 Pro zero-shot, val 60.2%
Olmo 3.1 32B Think val 60%
Gemma 4 12B val 60%
GPT-5.6 Sol zero-shot, val 59.3%
GPT-5 zero-shot, val 58.8%
Reka Flash 3.1 58.7%
GLM-5.3-Flash zero-shot, val 58.6%
Cogito v1 Preview 70B 58.6%
GPT-5.2 zero-shot, val 58.5%
glm-4.7-flash val 58.5%
Llama 3.1-70B 58.4%
Qwen3.5-9B val 58.2%
Gemini 3.1 Flash-Lite zero-shot, val 58.1%
o3 zero-shot, val 58.1%
Llama 4 Scout 58.1%
Reka Flash 3 58%
EXAONE 4.0 32B val 58%
Claude Sonnet 4.6 zero-shot, val 57.7%
o4-mini zero-shot, val 57.5%
Gemma 3 27B 57.4%
GPT-5.5 zero-shot, val 57%
Gemma 4 26B A4B val 56.9%
GPT-4.1 zero-shot, val 56.9%
Nemotron-Cascade 2 30B-A3B val 56.9%
Llama 3.3 70B Instruct zero-shot, val 56.8%
Gemma 3 12B 56.7%
QwQ-32B 56.6%
Apertus 1.5 70B val 56.4%
OLMo 2 32B Instruct 56.2%
GLM-4.5-Air thinking 56.2%
Qwen3.5-35B-A3B-Base val 56.1%
Aya Expanse 8B val 56%
GPT-5 nano zero-shot, val 56%
Claude Haiku 4.5 zero-shot, val 55.9%
Grok 4.20 thinking, zero-shot, val 55.9%
Gemini 2.5 Pro Jun 2025 zero-shot, val 55.7%
gpt-oss-120b val 55.6%
Gemini 2.5 Flash Sep 2025 zero-shot, val 55.5%
Qwen3-14B 55.5%
Ministral 3 3B Instruct zero-shot, val 55.5%
GPT-5.6 Terra zero-shot, val 55.5%
Apertus 70B 55.4%
Gemma 3n E4B 55.4%
Qwen3-235B-A22B 55.4%
Gemma 3n E2B 55.2%
GPT-6 Sol zero-shot, val 55.2%
Gemini 3.5 Flash-Lite zero-shot, val 55.1%
Qwen3-Next-80B-A3B thinking 55.1%
Grok-2 Dec 2024 zero-shot, val 55%
Phi-4 55%
Apertus 70B 54.8%
Grok 3 zero-shot, val 54.8%
Mistral Medium 3.5 zero-shot, val 54.8%
Voxtral Small 54.6%
Qwen3-8B 54.5%
gpt-oss-20b high 54.5%
Grok-3 mini zero-shot, val 54.5%
Gemma 3 27B PT 54.3%
Ministral 3 3B Instruct val 54.2%
Grok 4.1 Fast zero-shot, val 54.1%
Aya Expanse 32B 54.1%
GLM-4.1V-9B-Thinking 54%
gpt-oss-20b medium 53.7%
Qwen3-Next-80B-A3B-Instruct 53.3%
Gemini 2.5 Flash-Lite Sep 2025 zero-shot, val 53.2%
Gemma 3 12B PT 53%
Ministral 3 14B val 52.8%
DeepSeek-R1-Distill-Llama-70B 52.8%
Gemma 4 E4B val 52.8%
GPT-5 mini zero-shot, val 52.7%
Llama-3.1-Tulu-3-8B 52.6%
GPT-5.6 Luna zero-shot, val 52.5%
Olmo 3.1 32B Instruct 52.3%
Command R7B 52.1%
Qwen3-4B-Thinking-2507 52.1%
Llama 3.1-405B 51.9%
Granite 4.0 H Small 51.8%
Trinity Mini 51.1%
Qwen1.5-14B-Chat 51%
Granite-3.0-8B-Instruct 50.8%
Qwen3.5-4B val 50.8%
Aya Expanse 8B 49.7%
Gemma 4 E2B val 49.5%
DeepSeek-R1-Distill-Qwen-32B 49.5%
Ministral 3 8B Instruct val 49.4%
Apertus 1.5 70B zero-shot, val 49.4%
Qwen3-4B 49.3%
Falcon3-10B-Instruct 49.1%
deepseek-r1-0528-qwen3-8b 49%
Gemma 2 27B 49%
Ministral 3 3B Reasoning 48.7%
Tiny Aya Water val 48.6%
Granite-3.0-8B-Base 48.1%
Llama 3.1-8B 48%
gpt-oss-20b low 48%
nvidia-nemotron-nano-9b-v2 48%
granite-4.0-micro 48%
Tiny Aya Global val 47.9%
Falcon3-7B-Instruct 47.8%
qwen3-4b-instruct-2507 47.4%
Mistral 7B v0.3 47.4%
Gemma 7B 47.3%
Gemma 2 2B IT 47.1%
Gemma 3n E4B 47.1%
Yi-1.5-34B 46.6%
DeepSeek-R1-Distill-Qwen-14B 45.9%
Apertus 1.5 8B zero-shot, val 45.8%
Qwen3-1.7B 45.7%
Mistral 7B v0.1 45.2%
Mistral-7B-Instruct-v0.2 44.9%
Mistral-7B-Instruct-v0.1 44.7%
Apertus 1.5 8B val 44.1%
Qwen1.5-14B 43.4%
Qwen3.5-2B val 42.7%
AFM-4.5B 42.5%
Gemma 3 1B 41.9%
Qwen3.5-0.8B val 41%
Phi-4-reasoning 40.4%
Gemma 3 1B 40.3%
Gemma 4 31B IT val 40.1%
Gemma 2 9B 39.3%
Apertus 8B Instruct 39.2%
Gemma 2B 38.9%
Gemma 3 1B PT 36.7%
Hunyuan-4B-Instruct 36.1%
Gemma 4 26B A4B val 36%
Gemma 4 12B val 34.8%
Falcon3-3B-Instruct 30.8%
Llama 3.2 1B 28%
Qwen3-0.6B 27.4%
LFM2.5-1.2B-Thinking 25.8%
Qwen3.5-9B-Base val 24.3%
LFM2.5-1.2B-Instruct 23.2%
DeepSeek-R1-Distill-Qwen-7B 20%
Phi-2 19.4%
Phi-4-mini-reasoning 19.2%
DeepSeek-R1-Distill-Qwen-1.5B 17.4%
LFM2-1.2B 16.9%
Gemma 3 270M 16.2%
SmolLM2 1.7B Instruct 15.7%
Gemma 4 E2B val 12.4%
Gemma 4 E4B val 11.9%
SmolLM2-1.7B 8.53%
Qwen3.5-4B-Base val 6.93%
LFM2.5-8B-A1B val 6.70%
Phi-4-Reasoning-plus 6.42%
Gemma 3 270M 6.31%
Qwen3.5-2B-Base val 6.16%
SmolLM2-360M 4.99%
Qwen3.5-0.8B-Base val 4.45%
SmolLM2-135M 3.55%

8 models have held number one on the official leaderboard since 2023

Lab claims count from the day the card came out, independent runs from the model’s release.

Mistral AI Qwen Meta OpenAI Google DeepMind Z.ai
Own leaderboard 8 leaders

Source: Official Angry Tweets leaderboard

Scores over time

All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On the benchmark’s own leaderboard Best so far
0% 20% 40% 60% 80% 100% 2024 2025 2026 2027 Near the ceiling (85%) GPT-6.2 Astra Gemini 2.5 Flash o3-mini Qwen1.5-14B-Chat 0% 20% 40% 60% 80% 100% 2024 2025 2026 2027 Near the ceiling (85%) GPT-6.2 Astra o3-mini

Source: Official Angry Tweets leaderboard

No lab model card reports it yet

We read the headline results of 540 launch posts from 57 labs, and none of them list this benchmark.

More in Multilingual

All 545 in Multilingual
  • MMMLU In use MMLU professionally translated into 14 languages. Multiple-choice accuracy. Claude Mythos Preview 92.7%
  • WMT24++ In use English-to-55-language translation of WMT24 test data. Scored with automatic metrics such as MetricX or xCOMET. K-EXAONE 90.5%
  • MMLU-ProX In use MMLU-Pro translated into up to 29 languages with expert review. Multiple-choice accuracy per language. Qwen3.7-Max 87%
  • CLIcK In use Models answer 1,995 multiple-choice questions on Korean language and culture. Questions come from official Korean exams and textbooks in eleven categories. Solar Open 2 90.7%
  • Global MMLU Lite 3.0 In use Models answer 400 human-translated MMLU questions per language in 23 languages. Version 3.0 adds Czech, Hungarian, Oriya, Slovak and Tajik and updates Italian. Accuracy. Claude Opus 5 95.7%
  • MMMB In use Models answer multiple-choice questions about images in six languages across 15 categories. The set has about 12,000 questions. LFM2.5-VL-3B 83%

Data audit

All 11 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 5 pass
  • Pass Independent runs Scores from the official leaderboard (596 entries).
  • Pass Models scored 149 models we track, plus 447 other entries on its leaderboard.
  • Pass Newest score The newest is from November 2026, for GPT-6.2 Astra.
  • Pass Frontier models 3 of the top 10 have a score, led by GPT-6.2 Astra (#1).
  • Pass Headroom The best is 67%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on euroeval.com: 168 rows for models we track and 447 other entries at the last read.
  • Pass Last read November 20th, 2026.
  • Pass Read worked The last read worked.
Consistency 1 pass 1 note
  • Note Lab claims No model has both a lab number and an independent run.
  • Pass Known issues No published audit found problems.

How we check