Benchmark Almanac

FoSent

Gemini 2.5 Pro Jun 2025 91.6% leads since June 2025
Not in model cards

Models classify sentiment in Faroese news articles. Two native speakers annotated articles from Portalurin and Dimmalætting.

EuroEval scores sentiment labels with the Matthews correlation coefficient.

Released
May 2024
Kind
Reading in many languages
Level
Everyday
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Language
Faroese
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
540
Entries on its own leaderboard, 147 models we track

Gemini 2.5 Pro leads with 91.6% Gemini 2.5 Pro leads with 91.6%

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.

540 entries. Scroll sideways for the rest.

  1. 91.6 Gemini 2.5 Pro
  2. 84.4 speakleash/Bielik-11B-v3.1-Instruct
  3. 83.1 Gemini 3 Pro
  4. 82.8 gemini/gemini-3-flash-preview#thinking
  5. 81.8 Gemini 3.7 Flash
  6. 81.7 deepseek/deepseek-flash#no-thinking
  7. 81.0 o3
  8. 79.6 gemini/gemini-3-flash-preview#no-thinking
  9. 78.6 Qwen/Qwen3.6-27B-FP8
  10. 78.5 NbAiLab/borealis2-31b-preview
  11. 78.0 GPT-6.2 Astra
  12. 77.8 Claude Sonnet 4.6
  13. 77.8 GPT-4.1
  14. 77.7 GPT-6 Astra
  15. 77.7 GPT-5 mini
  16. 77.5 GPT-5.6 Terra
  17. 76.4 openai/gpt-5.4-mini-2026-03-17#low
  18. 76.3 Gemma 4 31B IT
  19. 75.9 GPT-6 Sol
  20. 75.3 Gemini 3.1 Flash-Lite
  21. 74.8 speakleash/Bielik-11B-v3.0-Instruct
  22. 74.6 claude-sonnet-4-5-20250929#no-thinking
  23. 74.6 deepseek/deepseek-flash#low
  24. 74.3 claude-sonnet-4-5-20250929#thinking
  25. 74.0 openai/gpt-5.4-mini-2026-03-17#high
  26. 73.9 Gemini 3.6 Flash
  27. 73.8 Grok-2
  28. 73.5 openai/gpt-5.4-mini-2026-03-17#medium
  29. 73.2 GPT-6.1 Sol
  30. 73.0 openai/gpt-5.4-mini-2026-03-17#none
  31. 72.9 gpt-5-mini-2025-08-07#high
  32. 72.9 Qwen/Qwen3.5-35B-A3B-FP8
  33. 72.8 GLM-5.3-Flash
  34. 72.4 Gemma 4 12B
  35. 71.7 NbAiLab/borealis2-26b-a4b-preview
  36. 71.4 Gemma 4 26B A4B
  37. 71.4 gemini/gemini-2.0-flash-001
  38. 71.3 GPT-5.2
  39. 71.0 mistralai/Mistral-Large-3-675B-Instruct-2512
  40. 71.0 claude-3-7-sonnet-20250219#thinking
  41. 71.0 openai/gpt-5.4-nano-2026-03-17#medium
  42. 70.9 GPT-6.2 Sol
  43. 70.3 Gemini 3.5 Flash-Lite
  44. 69.6 o4-mini
  45. 68.7 Apertus 1.5 8B
  46. 68.6 gemini/gemini-2.0-flash-lite-001
  47. 68.3 gemini/gemini-2.5-flash#no-thinking
  48. 67.9 openai/gpt-5.4-nano-2026-03-17#low
  49. 67.8 GadflyII/GLM-4.7-Flash-NVFP4
  50. 67.0 gpt-5-2025-08-07#high
  51. 66.0 Claude 3.5 Haiku
  52. 65.9 danish-foundation-models/DFM-Mimir-v1.8
  53. 65.5 NbAiLab/borealis-12b
  54. 65.2 glm-4.7-flash
  55. 65.1 GPT-5
  56. 64.9 Qwen3.5-35B-A3B-Base
  57. 64.4 gpt-5-mini-2025-08-07#minimal
  58. 64.2 Mistral Medium 3.5
  59. 64.1 mistralai/Ministral-3-14B-Instruct-2512
  60. 63.9 Cogito v1 Preview 70B
  61. 63.4 Claude Haiku 4.5
  62. 63.3 Qwen3.5-9B
  63. 63.2 ollama_chat/gemma3:27b-it-fp16
  64. 62.8 NbAiLab/borealis-open-12b
  65. 62.7 Unbabel/Tower-Plus-9B
  66. 62.5 GPT-6 Luna
  67. 62.4 gpt-5-2025-08-07#minimal
  68. 62.3 claude-3-7-sonnet-20250219#no-thinking
  69. 62.2 NbAiLab/borealis-27b
  70. 62.1 Gemini 1.5 Flash
  71. 62.1 Qwen 3.8 27B
  72. 62.0 NbAiLab/borealis-27b-instruct-preview
  73. 61.9 NbAiLab/borealis2-e4b-preview
  74. 61.8 danish-foundation-models/DFM-Mimir-v1.5
  75. 61.8 NbAiLab/borealis-12b-instruct-preview
  76. 61.5 NbAiLab/borealis-open-27b
  77. 61.3 Grok 3
  78. 61.2 Gemma 3 27B
  79. 61.1 gemini/gemini-2.5-flash-lite-preview-06-17#thinking
  80. 61.1 Gemini 2.5 Flash-Lite
  81. 60.5 LumiOpen/Llama-Poro-2-70B-Instruct
  82. 60.2 Gemma 4 E4B
  83. 59.9 o3-mini
  84. 59.6 GPT-4.1 mini
  85. 59.5 GPT-5.6 Sol
  86. 58.8 deepcogito/cogito-v2-preview-llama-70B
  87. 58.7 meta-llama/Llama-3.1-405B-Instruct-FP8
  88. 58.7 Apertus 1.5 70B
  89. 58.5 ThatsGroes/gemma-2-27b-it-FP8-Dynamic
  90. 58.1 Llama 3.1-405B
  91. 57.8 google/gemma-2-27b-it
  92. 57.1 openai/gpt-5.4-2026-03-05#none
  93. 57.0 typesafe/jev-1.13.0
  94. 57.0 Gemini 2.5 Flash
  95. 56.5 Qwen/Qwen3-235B-A22B-Instruct-2507-FP8
  96. 56.5 Gemma 2 27B
  97. 56.4 zai-org/GLM-4.5-Air#no-thinking
  98. 56.4 gpt-oss-120b
  99. 56.3 mistralai/Mistral-Small-4-119B-2603
  100. 56.0 LumiOpen/Llama-Poro-2-70B-SFT
  101. 55.9 GPT-5 nano
  102. 55.8 Qwen/Qwen3-235B-A22B-Thinking-2507-FP8
  103. 55.7 Ministral 3 8B Instruct
  104. 55.4 Gemma 3 12B PT
  105. 55.2 NovaSky-AI/Sky-T1-32B-Preview
  106. 54.8 mistralai/Ministral-3-8B-Base-2512
  107. 54.6 google/gemma-2-9b-it
  108. 54.4 Qwen3-235B-A22B
  109. 54.1 Gemma 3 12B
  110. 54.1 Unbabel/Tower-Plus-72B
  111. 53.9 Qwen3.5-4B
  112. 53.3 Llama 4 Scout
  113. 53.3 HPLT/NorOLMo-13B
  114. 53.3 Grok 4.1 Fast
  115. 53.2 meta-llama/Llama-4-Scout-17B-16E
  116. 53.0 Gemma 3n E4B
  117. 52.8 deepcogito/cogito-v1-preview-qwen-32B
  118. 52.7 Llama 3.1-70B
  119. 52.6 Grok-3 mini
  120. 52.3 Granite 4.0 H Small
  121. 52.2 Gemma 3 27B PT
  122. 52.1 mistralai/Ministral-3-14B-Base-2512
  123. 51.9 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8
  124. 51.6 openai/gpt-5.4-nano-2026-03-17#none
  125. 51.2 gemini/gemini-2.5-flash-lite-preview-06-17#no-thinking
  126. 51.2 gemini/gemini-2.5-flash-lite#no-thinking
  127. 51.0 Phi-4
  128. 50.8 mistralai/Mistral-Small-24B-Instruct-2501
  129. 50.7 swiss-ai/Apertus-v1.1-4B
  130. 50.6 deepcogito/cogito-v1-preview-qwen-14B
  131. 50.4 Qwen3-Next-80B-A3B-Instruct
  132. 50.2 Nemotron-Cascade 2 30B-A3B
  133. 50.2 Gemma 3n E2B
  134. 50.1 Qwen/Qwen3-32B#no-thinking
  135. 49.8 Apertus 70B
  136. 49.4 bullerwins/Magistral-Small-2506-fp8
  137. 48.8 mistralai/Mistral-Nemo-Instruct-2407
  138. 48.2 meta-llama/Llama-3.1-8B-Instruct
  139. 47.7 Qwen3-Next-80B-A3B
  140. 47.7 QwQ-32B
  141. 47.6 allenai/OLMo-2-0325-32B
  142. 47.3 mistralai/Mistral-Small-3.1-24B-Instruct-2503
  143. 47.2 Aya Expanse 32B
  144. 47.2 danish-foundation-models/DFM-Mimir-v1.8-Mini
  145. 47.1 mistralai/Mistral-Small-3.2-24B-Instruct-2506
  146. 46.8 152334H/miqu-1-70b-sf
  147. 46.5 Llama-3.1-Tulu-3-8B
  148. 46.5 Gemma 3 4B
  149. 46.4 allenai/Llama-3.1-Tulu-3.1-8B
  150. 46.4 allenai/Llama-3.1-Tulu-3-8B-DPO
  151. 46.2 GLM-4.5-Air
  152. 46.1 mistralai/Mistral-Small-3.1-24B-Base-2503
  153. 46.0 ollama_chat/gemma3:4b-it-q4_K_M
  154. 46.0 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16
  155. 45.8 mistralai/Magistral-Small-2509
  156. 45.6 utter-project/EuroLLM-22B-Instruct-2512
  157. 45.6 ollama_chat/phi4:14b-q4_K_M
  158. 45.5 Qwen3.5-4B-Base
  159. 45.4 Qwen/Qwen3-30B-A3B#no-thinking
  160. 45.4 NbAiLab/borealis-4b-instruct-preview
  161. 44.9 Reka Flash 3
  162. 44.8 Voxtral Small
  163. 44.8 ollama_chat/gemma3:12b-it-fp16
  164. 44.7 Gemma 2 9B
  165. 44.5 Apertus 8B Instruct
  166. 44.4 Reka Flash 3.1
  167. 44.3 ollama_chat/hf.co/google/gemma-3-4b-it-qat-q4_0-gguf
  168. 44.0 Gemma 3 4B PT
  169. 43.8 ollama_chat/gemma3:27b-it-q4_K_M
  170. 43.7 domyn/Domyn-Small-v1.0
  171. 43.5 Qwen/Qwen3-14B#no-thinking
  172. 43.5 OLMo 2 32B Instruct
  173. 43.3 gpt-5-nano-2025-08-07#high
  174. 43.1 deepseek-r1-0528-qwen3-8b
  175. 42.8 openai/gpt-5.4-nano-2026-03-17#high
  176. 42.4 Gemma 7B
  177. 42.4 Ministral 3 14B
  178. 42.1 norallm/normistral-11b-thinking
  179. 42.1 GPT-5.6 Luna
  180. 41.8 mistralai/Ministral-3-8B-Reasoning-2512
  181. 41.3 danish-foundation-models/DFM-Mimir
  182. 41.1 Qwen/Qwen3-8B#no-thinking
  183. 40.7 Qwen3.5-9B-Base
  184. 40.3 norallm/normistral-11b-long
  185. 40.2 Unbabel/Tower-Plus-2B
  186. 40.1 baidu/ERNIE-4.5-21B-A3B-PT
  187. 40.1 deepcogito/cogito-v1-preview-llama-8B
  188. 40.0 GLM-4.1V-9B-Thinking
  189. 40.0 Qwen3-8B
  190. 39.8 NousResearch/DeepHermes-3-Llama-3-8B-Preview
  191. 39.8 swiss-ai/Apertus-v1.1-4B-Instruct
  192. 39.6 ibm-granite/granite-4.0-h-tiny
  193. 39.5 mistralai/Ministral-3-3B-Base-2512
  194. 39.4 utter-project/EuroLLM-22B-Instruct-Preview
  195. 39.3 speakleash/Bielik-11B-v2.6-Instruct
  196. 39.3 ollama_chat/gemma3:12b-it-q4_K_M
  197. 39.0 norallm/normistral-11b-warm
  198. 38.9 openGPT-X/Teuken-7B-instruct-v0.6
  199. 38.6 TrustLLMeu/baseline-7-8b_1t-tokens_llama
  200. 38.5 Qwen3-14B
  201. 38.4 speakleash/Bielik-11B-v2.3-Instruct
  202. 37.9 Gemma 4 E2B
  203. 37.7 AI-Sweden-Models/Llama-3-8B-instruct
  204. 37.2 DeepSeek-R1-Distill-Llama-70B
  205. 37.2 gpt-oss-20b
  206. 37.1 NbAiLab/borealis-4b
  207. 37.0 TrustLLMeu/baseline-7-8b_2-3t-tokens_llama
  208. 36.7 granite-4.0-micro
  209. 36.6 nvidia-nemotron-nano-9b-v2
  210. 36.5 utter-project/EuroLLM-22B-Preview
  211. 36.5 Qwen3.5-2B
  212. 36.3 meta-llama/Llama-2-70b-hf
  213. 36.3 Llama 3.1-8B
  214. 36.1 ai21labs/Jamba-v0.1
  215. 35.9 amalia-llm/AMALIA-9B-0626-DPO
  216. 35.9 NbAiLab/borealis-open-4b
  217. 35.7 utter-project/EuroLLM-9B-Instruct
  218. 35.6 Qwen1.5-14B-Chat
  219. 35.4 Qwen3-4B-Thinking-2507
  220. 34.8 Ministral 3 3B Instruct
  221. 34.7 GPT-4.1 nano
  222. 34.6 bineric/lynx-instruct-30b#thinking
  223. 34.6 allenai/Olmo-3.1-32B-Instruct-DPO
  224. 34.0 utter-project/EuroLLM-9B
  225. 34.0 Falcon3-10B-Instruct
  226. 33.1 amalia-llm/AMALIA-9B-0626-SFT
  227. 33.1 speakleash/Bielik-Minitron-7B-v3.0-Instruct
  228. 33.1 swiss-ai/Apertus-v1.1-1.5B
  229. 32.9 allenai/Llama-3.1-Tulu-3-8B-SFT
  230. 32.7 tiiuae/falcon-11B
  231. 32.5 qwen3-4b-instruct-2507
  232. 32.5 bineric/lynx-instruct-30b#no-thinking
  233. 32.4 Olmo 3.1 32B Instruct
  234. 32.2 Command R7B
  235. 32.0 senseable/WestLake-7B-v2
  236. 32.0 allenai/Olmo-3.1-32B-Instruct-SFT
  237. 31.7 mistralai/Voxtral-Mini-3B-2507
  238. 31.6 EXAONE 4.0 32B
  239. 31.5 Falcon3-7B-Instruct
  240. 31.4 upstage/SOLAR-10.7B-v1.0
  241. 31.1 Granite-3.0-8B-Instruct
  242. 31.0 meta-llama/Llama-3.2-3B-Instruct
  243. 31.0 Qwen/Qwen3-4B#no-thinking
  244. 31.0 Qwen3-4B
  245. 30.8 01-ai/Yi-1.5-9B
  246. 30.6 NbAiLab/borealis-1b-instruct-preview
  247. 30.5 mii-llm/nesso-4B
  248. 30.3 HuggingFaceTB/SmolLM3-3B
  249. 30.0 deepcogito/cogito-v1-preview-llama-3B
  250. 29.7 Aya Expanse 8B
  251. 29.6 ibm-granite/granite-34b-code-base
  252. 29.5 Qwen3-1.7B
  253. 29.1 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-Base-BF16
  254. 29.0 Olmo 3.1 32B Think
  255. 29.0 AFM-4.5B
  256. 28.6 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6
  257. 28.5 Yi-1.5-34B
  258. 28.1 mistralai/Ministral-8B-Instruct-2410
  259. 27.9 allenai/OLMo-2-1124-13B-Instruct
  260. 27.7 tiiuae/Falcon3-10B-Base
  261. 27.6 Qwen/Qwen3-1.7B#no-thinking
  262. 27.6 Granite-3.0-8B-Base
  263. 27.4 Ministral 3 3B Reasoning
  264. 27.3 deahmed/Qwen3.5-2B-da-task
  265. 27.3 OpenLLM-France/Luciole-23B-Instruct-1.1
  266. 26.7 Mistral 7B v0.3
  267. 26.7 NbAiLab/borealis-open-1b
  268. 26.6 speakleash/Bielik-4.5B-v3.0-Instruct
  269. 26.5 utter-project/EuroLLM-9B-Instruct-2512
  270. 26.3 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16#no-thinking
  271. 26.2 Aleph-Alpha/Kolibri-1
  272. 25.8 HuggingFaceTB/SmolLM3-3B#no-thinking
  273. 25.8 allenai/OLMo-2-1124-13B
  274. 25.5 Mistral 7B v0.1
  275. 25.4 NbAiLab/borealis2-e2b-preview
  276. 25.3 ibm-granite/granite-20b-code-base
  277. 25.2 ibm-granite/granite-34b-code-instruct
  278. 24.6 TildeAI/TildeOpen-30b
  279. 24.5 ibm-granite/granite-4.0-h-micro
  280. 24.4 syvai/claes-v1
  281. 24.4 allenai/OLMo-2-1124-7B-Instruct
  282. 24.1 LumiOpen/Viking-33B
  283. 24.0 arcee-ai/AFM-4.5B-Base
  284. 24.0 CohereForAI/aya-23-35B
  285. 23.9 Gemma 2B
  286. 23.8 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660
  287. 23.8 Mistral-7B-Instruct-v0.2
  288. 23.6 speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551
  289. 23.4 ibm-granite/granite-8b-code-base-4k
  290. 23.1 tiiuae/Falcon3-7B-Base
  291. 22.8 CohereLabs/tiny-aya-base
  292. 22.2 Gemma 3 1B PT
  293. 22.2 bineric/lynx-instruct-30b
  294. 22.1 NbAiLab/borealis-1b
  295. 21.9 TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894
  296. 21.8 syvai/claes-preview
  297. 21.7 google/gemma-2-2b
  298. 21.3 DeepSeek-R1-Distill-Qwen-32B
  299. 21.2 OpenLLM-France/Luciole-8B-Instruct-1.1
  300. 21.1 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f
  301. 21.1 Qwen1.5-14B
  302. 21.1 DeepSeek-R1-Distill-Qwen-14B
  303. 21.1 Tiny Aya Global
  304. 20.8 Qwen/Qwen2.5-1.5B-Instruct
  305. 20.8 allenai/Olmo-3-1125-32B
  306. 20.3 meta-llama/Llama-2-70b-chat-hf
  307. 20.2 swiss-ai/Apertus-v1.1-1.5B-Instruct
  308. 20.0 Gemma 2 2B IT
  309. 19.5 Qwen3.5-2B-Base
  310. 19.3 Tiny Aya Water
  311. 19.1 allenai/OLMo-2-1124-7B
  312. 19.1 deepseek-ai/DeepSeek-R1-Distill-Llama-8B
  313. 18.9 nvidia/NVIDIA-Nemotron-3-Nano-4B-FP8
  314. 18.9 meta-llama/Llama-2-13b-chat-hf
  315. 18.8 allenai/Olmo-3-1025-7B
  316. 18.8 allenai/Olmo-3-7B-Think
  317. 18.7 TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49
  318. 18.5 AI-Sweden-Models/gpt-sw3-6.7b-v2-translator
  319. 18.1 nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16
  320. 17.9 LiquidAI/LFM2-2.6B
  321. 17.9 infly/OpenCoder-8B-Instruct
  322. 17.8 BSC-LT/ALIA-40b
  323. 17.8 allenai/Olmo-3-7B-Instruct-SFT
  324. 17.8 DeepSeek-R1-Distill-Qwen-7B
  325. 17.7 utter-project/EuroLLM-22B-2512
  326. 17.6 ibm-granite/granite-8b-code-instruct-4k
  327. 17.5 swiss-ai/Apertus-8B-2509
  328. 16.9 allenai/Olmo-3-7B-Instruct-DPO
  329. 16.8 infly/OpenCoder-8B-Base
  330. 16.5 Qwen/Qwen2.5-Math-7B
  331. 16.3 MaLA-LM/emma-500-llama2-7b
  332. 16.1 Qwen3-0.6B
  333. 15.7 allenai/Olmo-3-7B-Instruct
  334. 15.5 TildeAI/TildeOpen-30b#slow-tokenizer
  335. 15.2 NorwAI/NorwAI-Mixtral-8x7B-one-third-epoch
  336. 15.2 openeurollm/oellm-9b-128k-theta32m-v3
  337. 15.1 convaiinnovations/laya-multilingual
  338. 15.0 Qwen3.5-0.8B
  339. 14.9 OpenLLM-France/Luciole-8B-Base
  340. 14.2 BSC-LT/salamandra-7b-instruct
  341. 14.2 ibm-granite/granite-20b-code-instruct
  342. 13.5 speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c
  343. 13.4 mideind/IceBERT-xlmr-ic3
  344. 13.2 openGPT-X/Teuken-7B-instruct-commercial-v0.4
  345. 12.9 norallm/normistral-7b-scratch
  346. 12.8 AI-Sweden-Models/ModernBERT-large
  347. 12.8 Gemma 3 1B
  348. 12.5 LFM2-1.2B
  349. 12.5 common-pile/comma-v0.1-2t
  350. 12.2 google/gemma-7b-it
  351. 12.0 LumiOpen/Viking-13B
  352. 11.9 intfloat/multilingual-e5-large
  353. 11.9 DiscoResearch/DiscoLM-70b
  354. 11.5 OpenLLM-France/Luciole-1B-Instruct-1.1
  355. 11.4 Mistral-7B-Instruct-v0.1
  356. 11.3 yulan-team/YuLan-Mini-Instruct
  357. 10.8 EuroBERT/EuroBERT-210m
  358. 10.8 intfloat/multilingual-e5-base
  359. 10.7 Qwen/Qwen3-0.6B#no-thinking
  360. 10.7 HPLT/hplt2c_dan_checkpoints
  361. 10.5 jhu-clsp/mmBERT-base
  362. 10.5 arcee-ai/Trinity-Nano-Preview
  363. 10.4 occiglot/occiglot-7b-eu5
  364. 10.3 timpal0l/Mistral-7B-v0.1-flashback-v2
  365. 10.3 arcee-ai/Trinity-Mini-Base
  366. 10.2 meta-llama/Llama-2-13b-hf
  367. 10.2 utter-project/EuroVLM-1.7B-Preview
  368. 10.2 Phi-4-reasoning
  369. 9.79 gpt-5-nano-2025-08-07#minimal
  370. 9.66 meta-llama/Llama-2-7b-chat-hf
  371. 9.54 LiquidAI/LFM2-700M
  372. 9.47 ltg/norbert3-base
  373. 9.36 tencent/Hunyuan-7B-Instruct
  374. 9.03 LiquidAI/LFM2.5-8B-A1B-Base
  375. 8.54 EuroBERT/EuroBERT-610m
  376. 8.40 AI-Sweden-Models/gpt-sw3-20b
  377. 7.94 ibm-granite/granite-3.0-3b-a800m-instruct
  378. 7.85 NbAiLab/borealis-270m-instruct-preview
  379. 7.84 tencent/Hunyuan-1.8B-Instruct
  380. 7.66 marksverdhei/modern-norbert3-large
  381. 7.21 NbAiLab/borealis-open-270m
  382. 7.20 HuggingFaceTB/SmolLM3-3B-Base
  383. 7.20 swiss-ai/Apertus-v1.1-0.5B-Instruct
  384. 7.10 utter-project/EuroLLM-1.7B-Instruct
  385. 7.07 TurkuNLP/finnish-modernbert-large-seq-len-1024-117300-annealed
  386. 7.00 LFM2.5-1.2B-Thinking
  387. 6.91 swiss-ai/Apertus-v1.1-0.5B
  388. 6.56 ltg/norbert3-large
  389. 6.28 KBLab/roberta-base-swedish-cased
  390. 6.25 jhu-clsp/mmBERT-small
  391. 6.11 HPLT/hplt2c_isl_checkpoints
  392. 6.08 LFM2.5-8B-A1B
  393. 5.99 TurkuNLP/finnish-modernbert-tiny
  394. 5.95 common-pile/comma-v0.1-1t
  395. 5.78 SmolLM2 1.7B Instruct
  396. 5.54 NbAiLab/nb-roberta-base-ncc
  397. 5.53 HPLT/hplt2c_swe_checkpoints
  398. 5.38 ibm-granite/granite-3.0-3b-a800m-base
  399. 5.33 mideind/IceBERT-ic3
  400. 5.29 fresh-xlm-roberta-base
  401. 5.27 google-bert/bert-base-multilingual-uncased
  402. 5.25 vesteinn/DanskBERT
  403. 5.21 AI-Sweden-Models/gpt-sw3-126m
  404. 5.13 Qwen3.5-0.8B-Base
  405. 5.04 Gemma 3 270M
  406. 5.03 neurocode/IsRoBERTa
  407. 4.85 Llama 3.2 1B
  408. 4.76 occiglot/occiglot-7b-eu5-instruct
  409. 4.70 speakleash/Bielik-1.5B-v3.0-Instruct
  410. 4.69 answerdotai/ModernBERT-large
  411. 4.68 meta-llama/Llama-3.2-1B-Instruct
  412. 4.54 NbAiLab/nb-bert-base
  413. 4.53 Maltehb/aelaectra-danish-electra-small-cased
  414. 4.50 TurkuNLP/finnish-modernbert-base
  415. 4.46 mideind/IceBERT-large
  416. 4.33 facebook/MobileLLM-R1-360M-base
  417. 4.23 answerdotai/ModernBERT-base
  418. 4.15 vesteinn/FoBERT
  419. 4.02 TurkuNLP/finnish-modernbert-large
  420. 3.97 kyutai/helium-1-2b
  421. 3.89 AI-Sweden-Models/roberta-large-1160k
  422. 3.87 Falcon3-3B-Instruct
  423. 3.83 fastino/GLiNER2.5-Decide
  424. 3.74 NbAiLab/nb-bert-large
  425. 3.73 openbmb/MiniCPM5-1B
  426. 3.71 NbAiLab/nb-roberta-base-ncc-plus-scandi-1e4
  427. 3.63 arcee-ai/Trinity-Nano-Base
  428. 3.58 OpenLLM-France/Luciole-23B-Base
  429. 3.55 baidu/ERNIE-4.5-0.3B-PT
  430. 3.46 speakleash/Bielik-11B-v3-Base-20250730
  431. 3.39 vesteinn/XLMR-ENIS
  432. 3.37 mideind/IceBERT-mC4-is
  433. 3.33 OmniJev/OneJev1.6-0.8B
  434. 3.29 ltg/norbert3-small
  435. 3.24 AI-Sweden-Models/bert-large-nordic-pile-1M-steps
  436. 3.24 norallm/norbloom-7b-scratch
  437. 3.18 KBLab/bert-base-swedish-cased
  438. 3.06 NbAiLab/notram-bert-norwegian-cased-080321
  439. 3.04 DeepSeek-R1-Distill-Qwen-1.5B
  440. 2.99 ibm-granite/granite-3.0-1b-a400m-base
  441. 2.99 nvidia/NVIDIA-Nemotron-Nano-9B-v2#no-thinking
  442. 2.97 SmolLM2-360M
  443. 2.91 HPLT/gpt-13b-nordic-prerelease
  444. 2.82 NbAiLab/nb-roberta-base-ncc-plus-scandi-2e4
  445. 2.78 jaredpalmer/kev-0.5b
  446. 2.72 FacebookAI/xlm-roberta-large
  447. 2.69 vesteinn/ScandiBERT
  448. 2.63 KBLab/bert-base-swedish-cased-alpha
  449. 2.58 microsoft/mdeberta-v3-base
  450. 2.53 HPLT/hplt2c_nob_checkpoints
  451. 2.36 DDSC/roberta-base-danish
  452. 2.32 Phi-4-Reasoning-plus
  453. 2.31 danish-foundation-models/encoder-small-v1
  454. 2.29 google/gemma-2b-it
  455. 2.29 HuggingFaceTB/SmolLM2-360M-Instruct
  456. 2.28 KBLab/bert-base-swedish-cased-new
  457. 2.28 Maltehb/danish-bert-botxo
  458. 2.27 PleIAs/Pleias-1.2b-Preview
  459. 2.17 mideind/IceBERT-igc
  460. 2.13 AI-Sweden-Models/gpt-sw3-356m-instruct
  461. 2.06 KBLab/sentence-bert-swedish-cased
  462. 1.99 Muse-research/Muse-3B
  463. 1.97 NbAiLab/nb-roberta-base-scandi
  464. 1.96 utter-project/EuroMoE-2.6B-A0.6B-Preview
  465. 1.94 DGurgurov/xlm-r_dan-latn
  466. 1.82 NbAiLab/nb-roberta-base-scandi-1e4
  467. 1.73 DDSC/roberta-base-scandinavian
  468. 1.72 EuropeanParliament/EUBERT
  469. 1.63 LiquidAI/LFM2.5-1.2B-Base
  470. 1.59 Geotrend/distilbert-base-25lang-cased
  471. 1.54 AI-Sweden-Models/gpt-sw3-40b
  472. 1.40 facebook/MobileLLM-R1-140M-base
  473. 1.39 Qwen/Qwen2.5-Math-1.5B
  474. 1.38 SmolLM2-135M
  475. 1.37 AIDC-AI/Marco-Mini-Global-Base
  476. 1.24 OpenLLM-France/Luciole-1B-Base
  477. 1.23 infly/OpenCoder-1.5B-Base
  478. 1.12 mideind/IceBERT
  479. 1.05 LiquidAI/LFM2-350M
  480. 1.05 Qwen/Qwen1.5-0.5B
  481. 1.04 AI-Sweden-Models/gpt-sw3-126m-instruct
  482. 0.84 KBLab/megatron-bert-base-swedish-cased-125k
  483. 0.81 Qwen/Qwen1.5-0.5B-Chat
  484. 0.78 convaiinnovations/laya
  485. 0.77 SmolLM2-1.7B
  486. 0.69 HuggingFaceTB/SmolLM2-135M-Instruct
  487. 0.68 PleIAs/Baguettotron@refs/pr/6#no-thinking
  488. 0.66 Phi-4-mini-reasoning
  489. 0.64 vesteinn/ScandiBERT-no-faroese
  490. 0.44 ibm-granite/granite-3b-code-instruct-2k
  491. 0.34 ibm-granite/granite-3b-code-base-2k
  492. 0.34 marksverdhei/modern-norbert3-small
  493. 0.22 PleIAs/Pleias-350m-Preview
  494. 0.21 KBLab/megatron-bert-large-swedish-cased-165k
  495. 0.19 CohereLabs/aya-23-8B
  496. 0.19 infly/OpenCoder-1.5B-Instruct
  497. 0.19 ltg/norbert3-xs
  498. 0.18 NbAiLab/nb-bert-ncc-male2female
  499. 0.13 AI-Sweden-Models/gpt-sw3-356m
  500. 0.06 facebook/MobileLLM-R1-140M
  501. 0.05 Maltehb/aelaectra-danish-electra-small-uncased
  502. 0.04 marksverdhei/modern-norbert3-base
  503. 0.0 AI-Sweden-Models/gpt-sw3-1.3b-instruct
  504. 0.0 AI-Sweden-Models/gpt-sw3-6.7b-v2-instruct
  505. 0.0 neuralmagic/Sparse-Llama-3.1-8B-2of4
  506. 0.0 NorwAI/NorwAI-Magistral-24B-reasoning
  507. 0.0 openGPT-X/Teuken-7B-base-v0.6
  508. -0.05 NbAiLab/roberta_jan_128_ncc
  509. -0.08 openai-community/gpt2-medium
  510. -0.08 Phi-2
  511. -0.09 NbAiLab/borealis-270m
  512. -0.10 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview
  513. -0.29 ibm-granite/granite-3.0-1b-a400m-instruct
  514. -0.33 facebook/MobileLLM-R1-950M-base
  515. -0.47 KBLab/megatron-bert-base-swedish-cased-600k
  516. -0.58 marksverdhei/modern-norbert3-xs
  517. -0.61 openai-community/gpt2
  518. -0.68 Shekswess/trlm-135m
  519. -0.81 openai-community/gpt2-large
  520. -0.86 facebook/MobileLLM-R1-360M
  521. -1.03 fresh-electra-small
  522. -1.05 inclusionAI/Ling-mini-2.0
  523. -1.39 LFM2.5-1.2B-Instruct
  524. -1.42 AI-Sweden-Models/gpt-sw3-6.7b
  525. -1.42 AI-Sweden-Models/gpt-sw3-6.7b-v2
  526. -1.92 ltg/norbert2
  527. -2.11 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct
  528. -2.22 AI-Sweden-Models/gpt-sw3-1.3b
  529. -2.30 ltg/norbert
  530. -2.34 distilbert/distilbert-base-multilingual-cased
  531. -2.70 HPLT/hplt2c_nno_checkpoints
  532. -2.72 utter-project/EuroLLM-1.7B
  533. -3.26 BSC-LT/salamandra-7b
  534. -3.57 google-bert/bert-base-multilingual-cased
  535. -3.57 tiiuae/Falcon3-3B-Base
  536. -4.18 PleIAs/Monad
  537. -4.31 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base
  538. -5.17 inclusionAI/Ling-mini-base-2.0
  539. -5.80 norallm/normistral-11b-translate
  540. -7.13 facebook/MobileLLM-R1-950M
  1. 1 Gemini 2.5 Pro Jun 2025, zero-shot, val 91.6%
  2. 2 speakleash/Bielik-11B-v3.1-Instruct val 84.4%
  3. 3 Gemini 3 Pro zero-shot, val 83.1%
  4. 4 gemini/gemini-3-flash-preview#thinking zero-shot, val 82.8%
  5. 5 Gemini 3.7 Flash zero-shot, val 81.8%
  6. 6 deepseek/deepseek-flash#no-thinking val 81.7%
  7. 7 o3 zero-shot, val 81%
  8. 8 gemini/gemini-3-flash-preview#no-thinking zero-shot, val 79.6%
  9. 9 Qwen/Qwen3.6-27B-FP8 val 78.6%
  10. 10 NbAiLab/borealis2-31b-preview val 78.5%
  11. 11 GPT-6.2 Astra zero-shot, val 78%
  12. 12 Claude Sonnet 4.6 zero-shot, val 77.8%
  13. 13 GPT-4.1 zero-shot, val 77.8%
  14. 14 GPT-6 Astra zero-shot, val 77.7%
  15. 15 GPT-5 mini zero-shot, val 77.7%
  16. 16 GPT-5.6 Terra zero-shot, val 77.5%
  17. 17 openai/gpt-5.4-mini-2026-03-17#low zero-shot, val 76.4%
  18. 18 Gemma 4 31B IT val 76.3%
  19. 19 GPT-6 Sol zero-shot, val 75.9%
  20. 20 Gemini 3.1 Flash-Lite zero-shot, val 75.3%
  21. 21 speakleash/Bielik-11B-v3.0-Instruct val 74.8%
  22. 22 claude-sonnet-4-5-20250929#no-thinking zero-shot, val 74.6%
  23. 23 deepseek/deepseek-flash#low zero-shot, val 74.6%
  24. 24 claude-sonnet-4-5-20250929#thinking zero-shot, val 74.3%
  25. 25 openai/gpt-5.4-mini-2026-03-17#high zero-shot, val 74%
  26. 26 Gemini 3.6 Flash zero-shot, val 73.9%
  27. 27 Grok-2 Dec 2024, zero-shot, val 73.8%
  28. 28 openai/gpt-5.4-mini-2026-03-17#medium zero-shot, val 73.5%
  29. 29 GPT-6.1 Sol zero-shot, val 73.2%
  30. 30 openai/gpt-5.4-mini-2026-03-17#none zero-shot, val 73%
  31. 31 gpt-5-mini-2025-08-07#high zero-shot, val 72.9%
  32. 32 Qwen/Qwen3.5-35B-A3B-FP8 val 72.9%
  33. 33 GLM-5.3-Flash val 72.8%
  34. 34 Gemma 4 12B val 72.4%
  35. 35 NbAiLab/borealis2-26b-a4b-preview val 71.7%
  36. 36 Gemma 4 26B A4B val 71.4%
  37. 37 gemini/gemini-2.0-flash-001 zero-shot, val 71.4%
  38. 38 GPT-5.2 zero-shot, val 71.3%
  39. 39 mistralai/Mistral-Large-3-675B-Instruct-2512 val 71%
  40. 40 claude-3-7-sonnet-20250219#thinking zero-shot, val 71%
  41. 41 openai/gpt-5.4-nano-2026-03-17#medium zero-shot, val 71%
  42. 42 GPT-6.2 Sol zero-shot, val 70.9%
  43. 43 Gemini 3.5 Flash-Lite zero-shot, val 70.3%
  44. 44 o4-mini zero-shot, val 69.6%
  45. 45 Apertus 1.5 8B val 68.7%
  46. 46 gemini/gemini-2.0-flash-lite-001 zero-shot, val 68.6%
  47. 47 gemini/gemini-2.5-flash#no-thinking zero-shot, val 68.3%
  48. 48 openai/gpt-5.4-nano-2026-03-17#low zero-shot, val 67.9%
  49. 49 GadflyII/GLM-4.7-Flash-NVFP4 val 67.8%
  50. 50 gpt-5-2025-08-07#high zero-shot, val 67%
  51. 51 Claude 3.5 Haiku zero-shot, val 66%
  52. 52 danish-foundation-models/DFM-Mimir-v1.8 val 65.9%
  53. 53 NbAiLab/borealis-12b val 65.5%
  54. 54 glm-4.7-flash val 65.2%
  55. 55 GPT-5 zero-shot, val 65.1%
  56. 56 Qwen3.5-35B-A3B-Base val 64.9%
  57. 57 gpt-5-mini-2025-08-07#minimal zero-shot, val 64.4%
  58. 58 Mistral Medium 3.5 val 64.2%
  59. 59 mistralai/Ministral-3-14B-Instruct-2512 zero-shot, val 64.1%
  60. 60 Cogito v1 Preview 70B 63.9%
  61. 61 Claude Haiku 4.5 zero-shot, val 63.4%
  62. 62 Qwen3.5-9B val 63.3%
  63. 63 ollama_chat/gemma3:27b-it-fp16 63.2%
  64. 64 NbAiLab/borealis-open-12b val 62.8%
  65. 65 Unbabel/Tower-Plus-9B val 62.7%
  66. 66 GPT-6 Luna zero-shot, val 62.5%
  67. 67 gpt-5-2025-08-07#minimal zero-shot, val 62.4%
  68. 68 claude-3-7-sonnet-20250219#no-thinking zero-shot, val 62.3%
  69. 69 NbAiLab/borealis-27b val 62.2%
  70. 70 Gemini 1.5 Flash Sep 2024, zero-shot, val 62.1%
  71. 71 Qwen 3.8 27B val 62.1%
  72. 72 NbAiLab/borealis-27b-instruct-preview val 62%
  73. 73 NbAiLab/borealis2-e4b-preview val 61.9%
  74. 74 danish-foundation-models/DFM-Mimir-v1.5 zero-shot, val 61.8%
  75. 75 NbAiLab/borealis-12b-instruct-preview val 61.8%
  76. 76 NbAiLab/borealis-open-27b val 61.5%
  77. 77 Grok 3 zero-shot, val 61.3%
  78. 78 Gemma 3 27B zero-shot 61.2%
  79. 79 gemini/gemini-2.5-flash-lite-preview-06-17#thinking zero-shot, val 61.1%
  80. 80 Gemini 2.5 Flash-Lite Sep 2025, zero-shot, val 61.1%
  81. 81 LumiOpen/Llama-Poro-2-70B-Instruct 60.5%
  82. 82 Gemma 4 E4B val 60.2%
  83. 83 o3-mini zero-shot, val 59.9%
  84. 84 GPT-4.1 mini zero-shot, val 59.6%
  85. 85 GPT-5.6 Sol zero-shot, val 59.5%
  86. 86 deepcogito/cogito-v2-preview-llama-70B 58.8%
  87. 87 meta-llama/Llama-3.1-405B-Instruct-FP8 58.7%
  88. 88 Apertus 1.5 70B val 58.7%
  89. 89 ThatsGroes/gemma-2-27b-it-FP8-Dynamic 58.5%
  90. 90 Llama 3.1-405B 58.1%
  91. 91 google/gemma-2-27b-it 57.8%
  92. 92 openai/gpt-5.4-2026-03-05#none zero-shot, val 57.1%
  93. 93 typesafe/jev-1.13.0 zero-shot 57%
  94. 94 Gemini 2.5 Flash Sep 2025, zero-shot, val 57%
  95. 95 Qwen/Qwen3-235B-A22B-Instruct-2507-FP8 56.5%
  96. 96 Gemma 2 27B 56.5%
  97. 97 zai-org/GLM-4.5-Air#no-thinking 56.4%
  98. 98 gpt-oss-120b val 56.4%
  99. 99 mistralai/Mistral-Small-4-119B-2603 val 56.3%
  100. 100 LumiOpen/Llama-Poro-2-70B-SFT 56%
  101. 101 GPT-5 nano zero-shot, val 55.9%
  102. 102 Qwen/Qwen3-235B-A22B-Thinking-2507-FP8 55.8%
  103. 103 Ministral 3 8B Instruct val 55.7%
  104. 104 Gemma 3 12B PT 55.4%
  105. 105 NovaSky-AI/Sky-T1-32B-Preview 55.2%
  106. 106 mistralai/Ministral-3-8B-Base-2512 val 54.8%
  107. 107 google/gemma-2-9b-it 54.6%
  108. 108 Qwen3-235B-A22B 54.4%
  109. 109 Gemma 3 12B 54.1%
  110. 110 Unbabel/Tower-Plus-72B val 54.1%
  111. 111 Qwen3.5-4B val 53.9%
  112. 112 Llama 4 Scout 53.3%
  113. 113 HPLT/NorOLMo-13B val 53.3%
  114. 114 Grok 4.1 Fast zero-shot, val 53.3%
  115. 115 meta-llama/Llama-4-Scout-17B-16E 53.2%
  116. 116 Gemma 3n E4B 53%
  117. 117 deepcogito/cogito-v1-preview-qwen-32B 52.8%
  118. 118 Llama 3.1-70B 52.7%
  119. 119 Grok-3 mini zero-shot, val 52.6%
  120. 120 Granite 4.0 H Small 52.3%
  121. 121 Gemma 3 27B PT 52.2%
  122. 122 mistralai/Ministral-3-14B-Base-2512 val 52.1%
  123. 123 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8 val 51.9%
  124. 124 openai/gpt-5.4-nano-2026-03-17#none zero-shot, val 51.6%
  125. 125 gemini/gemini-2.5-flash-lite-preview-06-17#no-thinking zero-shot, val 51.2%
  126. 126 gemini/gemini-2.5-flash-lite#no-thinking zero-shot, val 51.2%
  127. 127 Phi-4 51%
  128. 128 mistralai/Mistral-Small-24B-Instruct-2501 50.8%
  129. 129 swiss-ai/Apertus-v1.1-4B val 50.7%
  130. 130 deepcogito/cogito-v1-preview-qwen-14B 50.6%
  131. 131 Qwen3-Next-80B-A3B-Instruct 50.4%
  132. 132 Nemotron-Cascade 2 30B-A3B val 50.2%
  133. 133 Gemma 3n E2B 50.2%
  134. 134 Qwen/Qwen3-32B#no-thinking 50.1%
  135. 135 Apertus 70B 49.8%
  136. 136 bullerwins/Magistral-Small-2506-fp8 49.4%
  137. 137 mistralai/Mistral-Nemo-Instruct-2407 48.8%
  138. 138 meta-llama/Llama-3.1-8B-Instruct 48.2%
  139. 139 Qwen3-Next-80B-A3B thinking 47.7%
  140. 140 QwQ-32B 47.7%
  141. 141 allenai/OLMo-2-0325-32B 47.6%
  142. 142 mistralai/Mistral-Small-3.1-24B-Instruct-2503 47.3%
  143. 143 Aya Expanse 32B 47.2%
  144. 144 danish-foundation-models/DFM-Mimir-v1.8-Mini zero-shot, val 47.2%
  145. 145 mistralai/Mistral-Small-3.2-24B-Instruct-2506 47.1%
  146. 146 152334H/miqu-1-70b-sf 46.8%
  147. 147 Llama-3.1-Tulu-3-8B 46.5%
  148. 148 Gemma 3 4B 46.5%
  149. 149 allenai/Llama-3.1-Tulu-3.1-8B 46.4%
  150. 150 allenai/Llama-3.1-Tulu-3-8B-DPO 46.4%
  151. 151 GLM-4.5-Air thinking 46.2%
  152. 152 mistralai/Mistral-Small-3.1-24B-Base-2503 46.1%
  153. 153 ollama_chat/gemma3:4b-it-q4_K_M 46%
  154. 154 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 val 46%
  155. 155 mistralai/Magistral-Small-2509 45.8%
  156. 156 utter-project/EuroLLM-22B-Instruct-2512 45.6%
  157. 157 ollama_chat/phi4:14b-q4_K_M val 45.6%
  158. 158 Qwen3.5-4B-Base val 45.5%
  159. 159 Qwen/Qwen3-30B-A3B#no-thinking 45.4%
  160. 160 NbAiLab/borealis-4b-instruct-preview 45.4%
  161. 161 Reka Flash 3 44.9%
  162. 162 Voxtral Small 44.8%
  163. 163 ollama_chat/gemma3:12b-it-fp16 val 44.8%
  164. 164 Gemma 2 9B 44.7%
  165. 165 Apertus 8B Instruct 44.5%
  166. 166 Reka Flash 3.1 44.4%
  167. 167 ollama_chat/hf.co/google/gemma-3-4b-it-qat-q4_0-gguf 44.3%
  168. 168 Gemma 3 4B PT 44%
  169. 169 ollama_chat/gemma3:27b-it-q4_K_M val 43.8%
  170. 170 domyn/Domyn-Small-v1.0 val 43.7%
  171. 171 Qwen/Qwen3-14B#no-thinking 43.5%
  172. 172 OLMo 2 32B Instruct 43.5%
  173. 173 gpt-5-nano-2025-08-07#high zero-shot, val 43.3%
  174. 174 deepseek-r1-0528-qwen3-8b 43.1%
  175. 175 openai/gpt-5.4-nano-2026-03-17#high zero-shot, val 42.8%
  176. 176 Gemma 7B 42.4%
  177. 177 Ministral 3 14B val 42.4%
  178. 178 norallm/normistral-11b-thinking 42.1%
  179. 179 GPT-5.6 Luna zero-shot, val 42.1%
  180. 180 mistralai/Ministral-3-8B-Reasoning-2512 41.8%
  181. 181 danish-foundation-models/DFM-Mimir val 41.3%
  182. 182 Qwen/Qwen3-8B#no-thinking 41.1%
  183. 183 Qwen3.5-9B-Base val 40.7%
  184. 184 norallm/normistral-11b-long 40.3%
  185. 185 Unbabel/Tower-Plus-2B val 40.2%
  186. 186 baidu/ERNIE-4.5-21B-A3B-PT 40.1%
  187. 187 deepcogito/cogito-v1-preview-llama-8B 40.1%
  188. 188 GLM-4.1V-9B-Thinking 40%
  189. 189 Qwen3-8B 40%
  190. 190 NousResearch/DeepHermes-3-Llama-3-8B-Preview 39.8%
  191. 191 swiss-ai/Apertus-v1.1-4B-Instruct val 39.8%
  192. 192 ibm-granite/granite-4.0-h-tiny 39.6%
  193. 193 mistralai/Ministral-3-3B-Base-2512 39.5%
  194. 194 utter-project/EuroLLM-22B-Instruct-Preview 39.4%
  195. 195 speakleash/Bielik-11B-v2.6-Instruct 39.3%
  196. 196 ollama_chat/gemma3:12b-it-q4_K_M val 39.3%
  197. 197 norallm/normistral-11b-warm 39%
  198. 198 openGPT-X/Teuken-7B-instruct-v0.6 val 38.9%
  199. 199 TrustLLMeu/baseline-7-8b_1t-tokens_llama 38.6%
  200. 200 Qwen3-14B 38.5%
  201. 201 speakleash/Bielik-11B-v2.3-Instruct 38.4%
  202. 202 Gemma 4 E2B val 37.9%
  203. 203 AI-Sweden-Models/Llama-3-8B-instruct val 37.7%
  204. 204 DeepSeek-R1-Distill-Llama-70B 37.2%
  205. 205 gpt-oss-20b medium 37.2%
  206. 206 NbAiLab/borealis-4b val 37.1%
  207. 207 TrustLLMeu/baseline-7-8b_2-3t-tokens_llama 37%
  208. 208 granite-4.0-micro 36.7%
  209. 209 nvidia-nemotron-nano-9b-v2 36.6%
  210. 210 utter-project/EuroLLM-22B-Preview 36.5%
  211. 211 Qwen3.5-2B val 36.5%
  212. 212 meta-llama/Llama-2-70b-hf 36.3%
  213. 213 Llama 3.1-8B 36.3%
  214. 214 ai21labs/Jamba-v0.1 36.1%
  215. 215 amalia-llm/AMALIA-9B-0626-DPO zero-shot, val 35.9%
  216. 216 NbAiLab/borealis-open-4b val 35.9%
  217. 217 utter-project/EuroLLM-9B-Instruct 35.7%
  218. 218 Qwen1.5-14B-Chat 35.6%
  219. 219 Qwen3-4B-Thinking-2507 35.4%
  220. 220 Ministral 3 3B Instruct val 34.8%
  221. 221 GPT-4.1 nano zero-shot, val 34.7%
  222. 222 bineric/lynx-instruct-30b#thinking 34.6%
  223. 223 allenai/Olmo-3.1-32B-Instruct-DPO 34.6%
  224. 224 utter-project/EuroLLM-9B 34%
  225. 225 Falcon3-10B-Instruct 34%
  226. 226 amalia-llm/AMALIA-9B-0626-SFT zero-shot, val 33.1%
  227. 227 speakleash/Bielik-Minitron-7B-v3.0-Instruct val 33.1%
  228. 228 swiss-ai/Apertus-v1.1-1.5B val 33.1%
  229. 229 allenai/Llama-3.1-Tulu-3-8B-SFT 32.9%
  230. 230 tiiuae/falcon-11B 32.7%
  231. 231 qwen3-4b-instruct-2507 32.5%
  232. 232 bineric/lynx-instruct-30b#no-thinking 32.5%
  233. 233 Olmo 3.1 32B Instruct 32.4%
  234. 234 Command R7B 32.2%
  235. 235 senseable/WestLake-7B-v2 32%
  236. 236 allenai/Olmo-3.1-32B-Instruct-SFT 32%
  237. 237 mistralai/Voxtral-Mini-3B-2507 31.7%
  238. 238 EXAONE 4.0 32B val 31.6%
  239. 239 Falcon3-7B-Instruct 31.5%
  240. 240 upstage/SOLAR-10.7B-v1.0 31.4%
  241. 241 Granite-3.0-8B-Instruct 31.1%
  242. 242 meta-llama/Llama-3.2-3B-Instruct 31%
  243. 243 Qwen/Qwen3-4B#no-thinking 31%
  244. 244 Qwen3-4B 31%
  245. 245 01-ai/Yi-1.5-9B 30.8%
  246. 246 NbAiLab/borealis-1b-instruct-preview val 30.6%
  247. 247 mii-llm/nesso-4B 30.5%
  248. 248 HuggingFaceTB/SmolLM3-3B 30.3%
  249. 249 deepcogito/cogito-v1-preview-llama-3B 30%
  250. 250 Aya Expanse 8B 29.7%
  251. 251 ibm-granite/granite-34b-code-base 29.6%
  252. 252 Qwen3-1.7B 29.5%
  253. 253 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-Base-BF16 val 29.1%
  254. 254 Olmo 3.1 32B Think val 29%
  255. 255 AFM-4.5B 29%
  256. 256 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6 28.6%
  257. 257 Yi-1.5-34B 28.5%
  258. 258 mistralai/Ministral-8B-Instruct-2410 28.1%
  259. 259 allenai/OLMo-2-1124-13B-Instruct 27.9%
  260. 260 tiiuae/Falcon3-10B-Base 27.7%
  261. 261 Qwen/Qwen3-1.7B#no-thinking 27.6%
  262. 262 Granite-3.0-8B-Base 27.6%
  263. 263 Ministral 3 3B Reasoning 27.4%
  264. 264 deahmed/Qwen3.5-2B-da-task val 27.3%
  265. 265 OpenLLM-France/Luciole-23B-Instruct-1.1 val 27.3%
  266. 266 Mistral 7B v0.3 26.7%
  267. 267 NbAiLab/borealis-open-1b val 26.7%
  268. 268 speakleash/Bielik-4.5B-v3.0-Instruct 26.6%
  269. 269 utter-project/EuroLLM-9B-Instruct-2512 val 26.5%
  270. 270 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16#no-thinking 26.3%
  271. 271 Aleph-Alpha/Kolibri-1 val 26.2%
  272. 272 HuggingFaceTB/SmolLM3-3B#no-thinking 25.8%
  273. 273 allenai/OLMo-2-1124-13B 25.8%
  274. 274 Mistral 7B v0.1 25.5%
  275. 275 NbAiLab/borealis2-e2b-preview zero-shot, val 25.4%
  276. 276 ibm-granite/granite-20b-code-base 25.3%
  277. 277 ibm-granite/granite-34b-code-instruct 25.2%
  278. 278 TildeAI/TildeOpen-30b val 24.6%
  279. 279 ibm-granite/granite-4.0-h-micro 24.5%
  280. 280 syvai/claes-v1 zero-shot 24.4%
  281. 281 allenai/OLMo-2-1124-7B-Instruct 24.4%
  282. 282 LumiOpen/Viking-33B 24.1%
  283. 283 arcee-ai/AFM-4.5B-Base 24%
  284. 284 CohereForAI/aya-23-35B 24%
  285. 285 Gemma 2B 23.9%
  286. 286 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660 23.8%
  287. 287 Mistral-7B-Instruct-v0.2 23.8%
  288. 288 speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551 zero-shot 23.6%
  289. 289 ibm-granite/granite-8b-code-base-4k 23.4%
  290. 290 tiiuae/Falcon3-7B-Base 23.1%
  291. 291 CohereLabs/tiny-aya-base val 22.8%
  292. 292 Gemma 3 1B PT 22.2%
  293. 293 bineric/lynx-instruct-30b val 22.2%
  294. 294 NbAiLab/borealis-1b val 22.1%
  295. 295 TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894 21.9%
  296. 296 syvai/claes-preview zero-shot 21.8%
  297. 297 google/gemma-2-2b 21.7%
  298. 298 DeepSeek-R1-Distill-Qwen-32B 21.3%
  299. 299 OpenLLM-France/Luciole-8B-Instruct-1.1 val 21.2%
  300. 300 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f 21.1%
  301. 301 Qwen1.5-14B 21.1%
  302. 302 DeepSeek-R1-Distill-Qwen-14B 21.1%
  303. 303 Tiny Aya Global val 21.1%
  304. 304 Qwen/Qwen2.5-1.5B-Instruct 20.8%
  305. 305 allenai/Olmo-3-1125-32B 20.8%
  306. 306 meta-llama/Llama-2-70b-chat-hf 20.3%
  307. 307 swiss-ai/Apertus-v1.1-1.5B-Instruct val 20.2%
  308. 308 Gemma 2 2B IT 20%
  309. 309 Qwen3.5-2B-Base val 19.5%
  310. 310 Tiny Aya Water val 19.3%
  311. 311 allenai/OLMo-2-1124-7B 19.1%
  312. 312 deepseek-ai/DeepSeek-R1-Distill-Llama-8B 19.1%
  313. 313 nvidia/NVIDIA-Nemotron-3-Nano-4B-FP8 val 18.9%
  314. 314 meta-llama/Llama-2-13b-chat-hf 18.9%
  315. 315 allenai/Olmo-3-1025-7B 18.8%
  316. 316 allenai/Olmo-3-7B-Think 18.8%
  317. 317 TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49 18.7%
  318. 318 AI-Sweden-Models/gpt-sw3-6.7b-v2-translator val 18.5%
  319. 319 nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16 18.1%
  320. 320 LiquidAI/LFM2-2.6B 17.9%
  321. 321 infly/OpenCoder-8B-Instruct 17.9%
  322. 322 BSC-LT/ALIA-40b 17.8%
  323. 323 allenai/Olmo-3-7B-Instruct-SFT 17.8%
  324. 324 DeepSeek-R1-Distill-Qwen-7B 17.8%
  325. 325 utter-project/EuroLLM-22B-2512 17.7%
  326. 326 ibm-granite/granite-8b-code-instruct-4k 17.6%
  327. 327 swiss-ai/Apertus-8B-2509 17.5%
  328. 328 allenai/Olmo-3-7B-Instruct-DPO 16.9%
  329. 329 infly/OpenCoder-8B-Base 16.8%
  330. 330 Qwen/Qwen2.5-Math-7B 16.5%
  331. 331 MaLA-LM/emma-500-llama2-7b 16.3%
  332. 332 Qwen3-0.6B 16.1%
  333. 333 allenai/Olmo-3-7B-Instruct 15.7%
  334. 334 TildeAI/TildeOpen-30b#slow-tokenizer 15.5%
  335. 335 NorwAI/NorwAI-Mixtral-8x7B-one-third-epoch val 15.2%
  336. 336 openeurollm/oellm-9b-128k-theta32m-v3 val 15.2%
  337. 337 convaiinnovations/laya-multilingual zero-shot 15.1%
  338. 338 Qwen3.5-0.8B val 15%
  339. 339 OpenLLM-France/Luciole-8B-Base val 14.9%
  340. 340 BSC-LT/salamandra-7b-instruct 14.2%
  341. 341 ibm-granite/granite-20b-code-instruct 14.2%
  342. 342 speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c zero-shot 13.5%
  343. 343 mideind/IceBERT-xlmr-ic3 13.4%
  344. 344 openGPT-X/Teuken-7B-instruct-commercial-v0.4 13.2%
  345. 345 norallm/normistral-7b-scratch val 12.9%
  346. 346 AI-Sweden-Models/ModernBERT-large 12.8%
  347. 347 Gemma 3 1B 12.8%
  348. 348 LFM2-1.2B 12.5%
  349. 349 common-pile/comma-v0.1-2t 12.5%
  350. 350 google/gemma-7b-it 12.2%
  351. 351 LumiOpen/Viking-13B 12%
  352. 352 intfloat/multilingual-e5-large 11.9%
  353. 353 DiscoResearch/DiscoLM-70b 11.9%
  354. 354 OpenLLM-France/Luciole-1B-Instruct-1.1 val 11.5%
  355. 355 Mistral-7B-Instruct-v0.1 11.4%
  356. 356 yulan-team/YuLan-Mini-Instruct 11.3%
  357. 357 EuroBERT/EuroBERT-210m 10.8%
  358. 358 intfloat/multilingual-e5-base 10.8%
  359. 359 Qwen/Qwen3-0.6B#no-thinking 10.7%
  360. 360 HPLT/hplt2c_dan_checkpoints 10.7%
  361. 361 jhu-clsp/mmBERT-base 10.5%
  362. 362 arcee-ai/Trinity-Nano-Preview 10.5%
  363. 363 occiglot/occiglot-7b-eu5 10.4%
  364. 364 timpal0l/Mistral-7B-v0.1-flashback-v2 10.3%
  365. 365 arcee-ai/Trinity-Mini-Base 10.3%
  366. 366 meta-llama/Llama-2-13b-hf 10.2%
  367. 367 utter-project/EuroVLM-1.7B-Preview 10.2%
  368. 368 Phi-4-reasoning 10.2%
  369. 369 gpt-5-nano-2025-08-07#minimal zero-shot, val 9.79%
  370. 370 meta-llama/Llama-2-7b-chat-hf 9.66%
  371. 371 LiquidAI/LFM2-700M 9.54%
  372. 372 ltg/norbert3-base 9.47%
  373. 373 tencent/Hunyuan-7B-Instruct 9.36%
  374. 374 LiquidAI/LFM2.5-8B-A1B-Base val 9.03%
  375. 375 EuroBERT/EuroBERT-610m 8.54%
  376. 376 AI-Sweden-Models/gpt-sw3-20b 8.40%
  377. 377 ibm-granite/granite-3.0-3b-a800m-instruct 7.94%
  378. 378 NbAiLab/borealis-270m-instruct-preview val 7.85%
  379. 379 tencent/Hunyuan-1.8B-Instruct 7.84%
  380. 380 marksverdhei/modern-norbert3-large 7.66%
  381. 381 NbAiLab/borealis-open-270m val 7.21%
  382. 382 HuggingFaceTB/SmolLM3-3B-Base 7.20%
  383. 383 swiss-ai/Apertus-v1.1-0.5B-Instruct val 7.20%
  384. 384 utter-project/EuroLLM-1.7B-Instruct val 7.10%
  385. 385 TurkuNLP/finnish-modernbert-large-seq-len-1024-117300-annealed 7.07%
  386. 386 LFM2.5-1.2B-Thinking 7.00%
  387. 387 swiss-ai/Apertus-v1.1-0.5B val 6.91%
  388. 388 ltg/norbert3-large 6.56%
  389. 389 KBLab/roberta-base-swedish-cased 6.28%
  390. 390 jhu-clsp/mmBERT-small 6.25%
  391. 391 HPLT/hplt2c_isl_checkpoints 6.11%
  392. 392 LFM2.5-8B-A1B val 6.08%
  393. 393 TurkuNLP/finnish-modernbert-tiny 5.99%
  394. 394 common-pile/comma-v0.1-1t 5.95%
  395. 395 SmolLM2 1.7B Instruct 5.78%
  396. 396 NbAiLab/nb-roberta-base-ncc 5.54%
  397. 397 HPLT/hplt2c_swe_checkpoints 5.53%
  398. 398 ibm-granite/granite-3.0-3b-a800m-base 5.38%
  399. 399 mideind/IceBERT-ic3 5.33%
  400. 400 fresh-xlm-roberta-base 5.29%
  401. 401 google-bert/bert-base-multilingual-uncased val 5.27%
  402. 402 vesteinn/DanskBERT 5.25%
  403. 403 AI-Sweden-Models/gpt-sw3-126m 5.21%
  404. 404 Qwen3.5-0.8B-Base val 5.13%
  405. 405 Gemma 3 270M 5.04%
  406. 406 neurocode/IsRoBERTa 5.03%
  407. 407 Llama 3.2 1B 4.85%
  408. 408 occiglot/occiglot-7b-eu5-instruct 4.76%
  409. 409 speakleash/Bielik-1.5B-v3.0-Instruct 4.70%
  410. 410 answerdotai/ModernBERT-large 4.69%
  411. 411 meta-llama/Llama-3.2-1B-Instruct 4.68%
  412. 412 NbAiLab/nb-bert-base 4.54%
  413. 413 Maltehb/aelaectra-danish-electra-small-cased 4.53%
  414. 414 TurkuNLP/finnish-modernbert-base 4.50%
  415. 415 mideind/IceBERT-large 4.46%
  416. 416 facebook/MobileLLM-R1-360M-base 4.33%
  417. 417 answerdotai/ModernBERT-base 4.23%
  418. 418 vesteinn/FoBERT 4.15%
  419. 419 TurkuNLP/finnish-modernbert-large 4.02%
  420. 420 kyutai/helium-1-2b 3.97%
  421. 421 AI-Sweden-Models/roberta-large-1160k 3.89%
  422. 422 Falcon3-3B-Instruct 3.87%
  423. 423 fastino/GLiNER2.5-Decide zero-shot 3.83%
  424. 424 NbAiLab/nb-bert-large 3.74%
  425. 425 openbmb/MiniCPM5-1B val 3.73%
  426. 426 NbAiLab/nb-roberta-base-ncc-plus-scandi-1e4 3.71%
  427. 427 arcee-ai/Trinity-Nano-Base 3.63%
  428. 428 OpenLLM-France/Luciole-23B-Base val 3.58%
  429. 429 baidu/ERNIE-4.5-0.3B-PT 3.55%
  430. 430 speakleash/Bielik-11B-v3-Base-20250730 3.46%
  431. 431 vesteinn/XLMR-ENIS 3.39%
  432. 432 mideind/IceBERT-mC4-is 3.37%
  433. 433 OmniJev/OneJev1.6-0.8B zero-shot 3.33%
  434. 434 ltg/norbert3-small 3.29%
  435. 435 AI-Sweden-Models/bert-large-nordic-pile-1M-steps 3.24%
  436. 436 norallm/norbloom-7b-scratch val 3.24%
  437. 437 KBLab/bert-base-swedish-cased 3.18%
  438. 438 NbAiLab/notram-bert-norwegian-cased-080321 3.06%
  439. 439 DeepSeek-R1-Distill-Qwen-1.5B 3.04%
  440. 440 ibm-granite/granite-3.0-1b-a400m-base 2.99%
  441. 441 nvidia/NVIDIA-Nemotron-Nano-9B-v2#no-thinking 2.99%
  442. 442 SmolLM2-360M 2.97%
  443. 443 HPLT/gpt-13b-nordic-prerelease 2.91%
  444. 444 NbAiLab/nb-roberta-base-ncc-plus-scandi-2e4 2.82%
  445. 445 jaredpalmer/kev-0.5b zero-shot 2.78%
  446. 446 FacebookAI/xlm-roberta-large 2.72%
  447. 447 vesteinn/ScandiBERT 2.69%
  448. 448 KBLab/bert-base-swedish-cased-alpha 2.63%
  449. 449 microsoft/mdeberta-v3-base 2.58%
  450. 450 HPLT/hplt2c_nob_checkpoints 2.53%
  451. 451 DDSC/roberta-base-danish 2.36%
  452. 452 Phi-4-Reasoning-plus 2.32%
  453. 453 danish-foundation-models/encoder-small-v1 2.31%
  454. 454 google/gemma-2b-it 2.29%
  455. 455 HuggingFaceTB/SmolLM2-360M-Instruct 2.29%
  456. 456 KBLab/bert-base-swedish-cased-new 2.28%
  457. 457 Maltehb/danish-bert-botxo 2.28%
  458. 458 PleIAs/Pleias-1.2b-Preview 2.27%
  459. 459 mideind/IceBERT-igc 2.17%
  460. 460 AI-Sweden-Models/gpt-sw3-356m-instruct 2.13%
  461. 461 KBLab/sentence-bert-swedish-cased 2.06%
  462. 462 Muse-research/Muse-3B val 1.99%
  463. 463 NbAiLab/nb-roberta-base-scandi 1.97%
  464. 464 utter-project/EuroMoE-2.6B-A0.6B-Preview 1.96%
  465. 465 DGurgurov/xlm-r_dan-latn 1.94%
  466. 466 NbAiLab/nb-roberta-base-scandi-1e4 1.82%
  467. 467 DDSC/roberta-base-scandinavian 1.73%
  468. 468 EuropeanParliament/EUBERT 1.72%
  469. 469 LiquidAI/LFM2.5-1.2B-Base 1.63%
  470. 470 Geotrend/distilbert-base-25lang-cased 1.59%
  471. 471 AI-Sweden-Models/gpt-sw3-40b val 1.54%
  472. 472 facebook/MobileLLM-R1-140M-base 1.40%
  473. 473 Qwen/Qwen2.5-Math-1.5B 1.39%
  474. 474 SmolLM2-135M 1.38%
  475. 475 AIDC-AI/Marco-Mini-Global-Base val 1.37%
  476. 476 OpenLLM-France/Luciole-1B-Base val 1.24%
  477. 477 infly/OpenCoder-1.5B-Base 1.23%
  478. 478 mideind/IceBERT 1.12%
  479. 479 LiquidAI/LFM2-350M 1.05%
  480. 480 Qwen/Qwen1.5-0.5B 1.05%
  481. 481 AI-Sweden-Models/gpt-sw3-126m-instruct 1.04%
  482. 482 KBLab/megatron-bert-base-swedish-cased-125k 0.84%
  483. 483 Qwen/Qwen1.5-0.5B-Chat 0.81%
  484. 484 convaiinnovations/laya zero-shot 0.78%
  485. 485 SmolLM2-1.7B 0.77%
  486. 486 HuggingFaceTB/SmolLM2-135M-Instruct 0.69%
  487. 487 PleIAs/Baguettotron@refs/pr/6#no-thinking 0.68%
  488. 488 Phi-4-mini-reasoning 0.66%
  489. 489 vesteinn/ScandiBERT-no-faroese 0.64%
  490. 490 ibm-granite/granite-3b-code-instruct-2k 0.44%
  491. 491 ibm-granite/granite-3b-code-base-2k 0.34%
  492. 492 marksverdhei/modern-norbert3-small 0.34%
  493. 493 PleIAs/Pleias-350m-Preview 0.22%
  494. 494 KBLab/megatron-bert-large-swedish-cased-165k 0.21%
  495. 495 CohereLabs/aya-23-8B val 0.19%
  496. 496 infly/OpenCoder-1.5B-Instruct 0.19%
  497. 497 ltg/norbert3-xs 0.19%
  498. 498 NbAiLab/nb-bert-ncc-male2female 0.18%
  499. 499 AI-Sweden-Models/gpt-sw3-356m 0.13%
  500. 500 facebook/MobileLLM-R1-140M 0.06%
  501. 501 Maltehb/aelaectra-danish-electra-small-uncased 0.05%
  502. 502 marksverdhei/modern-norbert3-base 0.04%
  503. 503 AI-Sweden-Models/gpt-sw3-1.3b-instruct val 0%
  504. 504 AI-Sweden-Models/gpt-sw3-6.7b-v2-instruct val 0%
  505. 505 neuralmagic/Sparse-Llama-3.1-8B-2of4 0%
  506. 506 NorwAI/NorwAI-Magistral-24B-reasoning val 0%
  507. 507 openGPT-X/Teuken-7B-base-v0.6 val 0%
  508. 508 NbAiLab/roberta_jan_128_ncc -0.05%
  509. 509 openai-community/gpt2-medium -0.08%
  510. 510 Phi-2 -0.08%
  511. 511 NbAiLab/borealis-270m val -0.09%
  512. 512 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview -0.10%
  513. 513 ibm-granite/granite-3.0-1b-a400m-instruct -0.29%
  514. 514 facebook/MobileLLM-R1-950M-base -0.33%
  515. 515 KBLab/megatron-bert-base-swedish-cased-600k -0.47%
  516. 516 marksverdhei/modern-norbert3-xs -0.58%
  517. 517 openai-community/gpt2 -0.61%
  518. 518 Shekswess/trlm-135m -0.68%
  519. 519 openai-community/gpt2-large -0.81%
  520. 520 facebook/MobileLLM-R1-360M -0.86%
  521. 521 fresh-electra-small -1.03%
  522. 522 inclusionAI/Ling-mini-2.0 val -1.05%
  523. 523 LFM2.5-1.2B-Instruct -1.39%
  524. 524 AI-Sweden-Models/gpt-sw3-6.7b val -1.42%
  525. 525 AI-Sweden-Models/gpt-sw3-6.7b-v2 val -1.42%
  526. 526 ltg/norbert2 -1.92%
  527. 527 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct -2.11%
  528. 528 AI-Sweden-Models/gpt-sw3-1.3b val -2.22%
  529. 529 ltg/norbert -2.30%
  530. 530 distilbert/distilbert-base-multilingual-cased -2.34%
  531. 531 HPLT/hplt2c_nno_checkpoints -2.70%
  532. 532 utter-project/EuroLLM-1.7B val -2.72%
  533. 533 BSC-LT/salamandra-7b -3.26%
  534. 534 google-bert/bert-base-multilingual-cased val -3.57%
  535. 535 tiiuae/Falcon3-3B-Base -3.57%
  536. 536 PleIAs/Monad -4.18%
  537. 537 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base -4.31%
  538. 538 inclusionAI/Ling-mini-base-2.0 val -5.17%
  539. 539 norallm/normistral-11b-translate val -5.80%
  540. 540 facebook/MobileLLM-R1-950M -7.13%

Source: Official FoSent leaderboard

147 models. Scroll sideways for the rest.

  1. 91.6 Gemini 2.5 Pro
  2. 83.1 Gemini 3 Pro
  3. 81.8 Gemini 3.7 Flash
  4. 81.0 o3
  5. 78.0 GPT-6.2 Astra
  6. 77.8 Claude Sonnet 4.6
  7. 77.8 GPT-4.1
  8. 77.7 GPT-6 Astra
  9. 77.7 GPT-5 mini
  10. 77.5 GPT-5.6 Terra
  11. 76.3 Gemma 4 31B IT
  12. 75.9 GPT-6 Sol
  13. 75.3 Gemini 3.1 Flash-Lite
  14. 73.9 Gemini 3.6 Flash
  15. 73.8 Grok-2
  16. 73.2 GPT-6.1 Sol
  17. 72.8 GLM-5.3-Flash
  18. 72.4 Gemma 4 12B
  19. 71.4 Gemma 4 26B A4B
  20. 71.3 GPT-5.2
  21. 70.9 GPT-6.2 Sol
  22. 70.3 Gemini 3.5 Flash-Lite
  23. 69.6 o4-mini
  24. 68.7 Apertus 1.5 8B
  25. 66.0 Claude 3.5 Haiku
  26. 65.2 glm-4.7-flash
  27. 65.1 GPT-5
  28. 64.9 Qwen3.5-35B-A3B-Base
  29. 64.2 Mistral Medium 3.5
  30. 63.9 Cogito v1 Preview 70B
  31. 63.4 Claude Haiku 4.5
  32. 63.3 Qwen3.5-9B
  33. 62.5 GPT-6 Luna
  34. 62.1 Gemini 1.5 Flash
  35. 62.1 Qwen 3.8 27B
  36. 61.3 Grok 3
  37. 61.2 Gemma 3 27B
  38. 61.1 Gemini 2.5 Flash-Lite
  39. 60.2 Gemma 4 E4B
  40. 59.9 o3-mini
  41. 59.6 GPT-4.1 mini
  42. 59.5 GPT-5.6 Sol
  43. 58.7 Apertus 1.5 70B
  44. 58.1 Llama 3.1-405B
  45. 57.0 Gemini 2.5 Flash
  46. 56.5 Gemma 2 27B
  47. 56.4 gpt-oss-120b
  48. 55.9 GPT-5 nano
  49. 55.7 Ministral 3 8B Instruct
  50. 55.4 Gemma 3 12B PT
  51. 54.4 Qwen3-235B-A22B
  52. 54.1 Gemma 3 12B
  53. 53.9 Qwen3.5-4B
  54. 53.3 Llama 4 Scout
  55. 53.3 Grok 4.1 Fast
  56. 53.0 Gemma 3n E4B
  57. 52.7 Llama 3.1-70B
  58. 52.6 Grok-3 mini
  59. 52.3 Granite 4.0 H Small
  60. 52.2 Gemma 3 27B PT
  61. 51.0 Phi-4
  62. 50.4 Qwen3-Next-80B-A3B-Instruct
  63. 50.2 Nemotron-Cascade 2 30B-A3B
  64. 50.2 Gemma 3n E2B
  65. 49.8 Apertus 70B
  66. 47.7 Qwen3-Next-80B-A3B
  67. 47.7 QwQ-32B
  68. 47.2 Aya Expanse 32B
  69. 46.5 Llama-3.1-Tulu-3-8B
  70. 46.5 Gemma 3 4B
  71. 46.2 GLM-4.5-Air
  72. 45.5 Qwen3.5-4B-Base
  73. 44.9 Reka Flash 3
  74. 44.8 Voxtral Small
  75. 44.7 Gemma 2 9B
  76. 44.5 Apertus 8B Instruct
  77. 44.4 Reka Flash 3.1
  78. 44.0 Gemma 3 4B PT
  79. 43.5 OLMo 2 32B Instruct
  80. 43.1 deepseek-r1-0528-qwen3-8b
  81. 42.4 Gemma 7B
  82. 42.4 Ministral 3 14B
  83. 42.1 GPT-5.6 Luna
  84. 40.7 Qwen3.5-9B-Base
  85. 40.0 GLM-4.1V-9B-Thinking
  86. 40.0 Qwen3-8B
  87. 38.5 Qwen3-14B
  88. 37.9 Gemma 4 E2B
  89. 37.2 DeepSeek-R1-Distill-Llama-70B
  90. 37.2 gpt-oss-20b
  91. 36.7 granite-4.0-micro
  92. 36.6 nvidia-nemotron-nano-9b-v2
  93. 36.5 Qwen3.5-2B
  94. 36.3 Llama 3.1-8B
  95. 35.6 Qwen1.5-14B-Chat
  96. 35.4 Qwen3-4B-Thinking-2507
  97. 34.8 Ministral 3 3B Instruct
  98. 34.7 GPT-4.1 nano
  99. 34.0 Falcon3-10B-Instruct
  100. 32.5 qwen3-4b-instruct-2507
  101. 32.4 Olmo 3.1 32B Instruct
  102. 32.2 Command R7B
  103. 31.6 EXAONE 4.0 32B
  104. 31.5 Falcon3-7B-Instruct
  105. 31.1 Granite-3.0-8B-Instruct
  106. 31.0 Qwen3-4B
  107. 29.7 Aya Expanse 8B
  108. 29.5 Qwen3-1.7B
  109. 29.0 Olmo 3.1 32B Think
  110. 29.0 AFM-4.5B
  111. 28.5 Yi-1.5-34B
  112. 27.6 Granite-3.0-8B-Base
  113. 27.4 Ministral 3 3B Reasoning
  114. 26.7 Mistral 7B v0.3
  115. 25.5 Mistral 7B v0.1
  116. 23.9 Gemma 2B
  117. 23.8 Mistral-7B-Instruct-v0.2
  118. 22.2 Gemma 3 1B PT
  119. 21.3 DeepSeek-R1-Distill-Qwen-32B
  120. 21.1 Qwen1.5-14B
  121. 21.1 DeepSeek-R1-Distill-Qwen-14B
  122. 21.1 Tiny Aya Global
  123. 20.0 Gemma 2 2B IT
  124. 19.5 Qwen3.5-2B-Base
  125. 19.3 Tiny Aya Water
  126. 17.8 DeepSeek-R1-Distill-Qwen-7B
  127. 16.1 Qwen3-0.6B
  128. 15.0 Qwen3.5-0.8B
  129. 12.8 Gemma 3 1B
  130. 12.5 LFM2-1.2B
  131. 11.4 Mistral-7B-Instruct-v0.1
  132. 10.2 Phi-4-reasoning
  133. 7.00 LFM2.5-1.2B-Thinking
  134. 6.08 LFM2.5-8B-A1B
  135. 5.78 SmolLM2 1.7B Instruct
  136. 5.13 Qwen3.5-0.8B-Base
  137. 5.04 Gemma 3 270M
  138. 4.85 Llama 3.2 1B
  139. 3.87 Falcon3-3B-Instruct
  140. 3.04 DeepSeek-R1-Distill-Qwen-1.5B
  141. 2.97 SmolLM2-360M
  142. 2.32 Phi-4-Reasoning-plus
  143. 1.38 SmolLM2-135M
  144. 0.77 SmolLM2-1.7B
  145. 0.66 Phi-4-mini-reasoning
  146. -0.08 Phi-2
  147. -1.39 LFM2.5-1.2B-Instruct
  1. 1 Gemini 2.5 Pro Jun 2025, zero-shot, val 91.6%
  2. 2 Gemini 3 Pro zero-shot, val 83.1%
  3. 3 Gemini 3.7 Flash zero-shot, val 81.8%
  4. 4 o3 zero-shot, val 81%
  5. 5 GPT-6.2 Astra zero-shot, val 78%
  6. 6 Claude Sonnet 4.6 zero-shot, val 77.8%
  7. 7 GPT-4.1 zero-shot, val 77.8%
  8. 8 GPT-6 Astra zero-shot, val 77.7%
  9. 9 GPT-5 mini zero-shot, val 77.7%
  10. 10 GPT-5.6 Terra zero-shot, val 77.5%
  11. 11 Gemma 4 31B IT val 76.3%
  12. 12 GPT-6 Sol zero-shot, val 75.9%
  13. 13 Gemini 3.1 Flash-Lite zero-shot, val 75.3%
  14. 14 Gemini 3.6 Flash zero-shot, val 73.9%
  15. 15 Grok-2 Dec 2024, zero-shot, val 73.8%
  16. 16 GPT-6.1 Sol zero-shot, val 73.2%
  17. 17 GLM-5.3-Flash val 72.8%
  18. 18 Gemma 4 12B val 72.4%
  19. 19 Gemma 4 26B A4B val 71.4%
  20. 20 GPT-5.2 zero-shot, val 71.3%
  21. 21 GPT-6.2 Sol zero-shot, val 70.9%
  22. 22 Gemini 3.5 Flash-Lite zero-shot, val 70.3%
  23. 23 o4-mini zero-shot, val 69.6%
  24. 24 Apertus 1.5 8B val 68.7%
  25. 25 Claude 3.5 Haiku zero-shot, val 66%
  26. 26 glm-4.7-flash val 65.2%
  27. 27 GPT-5 zero-shot, val 65.1%
  28. 28 Qwen3.5-35B-A3B-Base val 64.9%
  29. 29 Mistral Medium 3.5 val 64.2%
  30. 30 Cogito v1 Preview 70B 63.9%
  31. 31 Claude Haiku 4.5 zero-shot, val 63.4%
  32. 32 Qwen3.5-9B val 63.3%
  33. 33 GPT-6 Luna zero-shot, val 62.5%
  34. 34 Gemini 1.5 Flash Sep 2024, zero-shot, val 62.1%
  35. 35 Qwen 3.8 27B val 62.1%
  36. 36 Grok 3 zero-shot, val 61.3%
  37. 37 Gemma 3 27B zero-shot 61.2%
  38. 38 Gemini 2.5 Flash-Lite Sep 2025, zero-shot, val 61.1%
  39. 39 Gemma 4 E4B val 60.2%
  40. 40 o3-mini zero-shot, val 59.9%
  41. 41 GPT-4.1 mini zero-shot, val 59.6%
  42. 42 GPT-5.6 Sol zero-shot, val 59.5%
  43. 43 Apertus 1.5 70B val 58.7%
  44. 44 Llama 3.1-405B 58.1%
  45. 45 Gemini 2.5 Flash Sep 2025, zero-shot, val 57%
  46. 46 Gemma 2 27B 56.5%
  47. 47 gpt-oss-120b val 56.4%
  48. 48 GPT-5 nano zero-shot, val 55.9%
  49. 49 Ministral 3 8B Instruct val 55.7%
  50. 50 Gemma 3 12B PT 55.4%
  51. 51 Qwen3-235B-A22B 54.4%
  52. 52 Gemma 3 12B 54.1%
  53. 53 Qwen3.5-4B val 53.9%
  54. 54 Llama 4 Scout 53.3%
  55. 55 Grok 4.1 Fast zero-shot, val 53.3%
  56. 56 Gemma 3n E4B 53%
  57. 57 Llama 3.1-70B 52.7%
  58. 58 Grok-3 mini zero-shot, val 52.6%
  59. 59 Granite 4.0 H Small 52.3%
  60. 60 Gemma 3 27B PT 52.2%
  61. 61 Phi-4 51%
  62. 62 Qwen3-Next-80B-A3B-Instruct 50.4%
  63. 63 Nemotron-Cascade 2 30B-A3B val 50.2%
  64. 64 Gemma 3n E2B 50.2%
  65. 65 Apertus 70B 49.8%
  66. 66 Qwen3-Next-80B-A3B thinking 47.7%
  67. 67 QwQ-32B 47.7%
  68. 68 Aya Expanse 32B 47.2%
  69. 69 Llama-3.1-Tulu-3-8B 46.5%
  70. 70 Gemma 3 4B 46.5%
  71. 71 GLM-4.5-Air thinking 46.2%
  72. 72 Qwen3.5-4B-Base val 45.5%
  73. 73 Reka Flash 3 44.9%
  74. 74 Voxtral Small 44.8%
  75. 75 Gemma 2 9B 44.7%
  76. 76 Apertus 8B Instruct 44.5%
  77. 77 Reka Flash 3.1 44.4%
  78. 78 Gemma 3 4B PT 44%
  79. 79 OLMo 2 32B Instruct 43.5%
  80. 80 deepseek-r1-0528-qwen3-8b 43.1%
  81. 81 Gemma 7B 42.4%
  82. 82 Ministral 3 14B val 42.4%
  83. 83 GPT-5.6 Luna zero-shot, val 42.1%
  84. 84 Qwen3.5-9B-Base val 40.7%
  85. 85 GLM-4.1V-9B-Thinking 40%
  86. 86 Qwen3-8B 40%
  87. 87 Qwen3-14B 38.5%
  88. 88 Gemma 4 E2B val 37.9%
  89. 89 DeepSeek-R1-Distill-Llama-70B 37.2%
  90. 90 gpt-oss-20b medium 37.2%
  91. 91 granite-4.0-micro 36.7%
  92. 92 nvidia-nemotron-nano-9b-v2 36.6%
  93. 93 Qwen3.5-2B val 36.5%
  94. 94 Llama 3.1-8B 36.3%
  95. 95 Qwen1.5-14B-Chat 35.6%
  96. 96 Qwen3-4B-Thinking-2507 35.4%
  97. 97 Ministral 3 3B Instruct val 34.8%
  98. 98 GPT-4.1 nano zero-shot, val 34.7%
  99. 99 Falcon3-10B-Instruct 34%
  100. 100 qwen3-4b-instruct-2507 32.5%
  101. 101 Olmo 3.1 32B Instruct 32.4%
  102. 102 Command R7B 32.2%
  103. 103 EXAONE 4.0 32B val 31.6%
  104. 104 Falcon3-7B-Instruct 31.5%
  105. 105 Granite-3.0-8B-Instruct 31.1%
  106. 106 Qwen3-4B 31%
  107. 107 Aya Expanse 8B 29.7%
  108. 108 Qwen3-1.7B 29.5%
  109. 109 Olmo 3.1 32B Think val 29%
  110. 110 AFM-4.5B 29%
  111. 111 Yi-1.5-34B 28.5%
  112. 112 Granite-3.0-8B-Base 27.6%
  113. 113 Ministral 3 3B Reasoning 27.4%
  114. 114 Mistral 7B v0.3 26.7%
  115. 115 Mistral 7B v0.1 25.5%
  116. 116 Gemma 2B 23.9%
  117. 117 Mistral-7B-Instruct-v0.2 23.8%
  118. 118 Gemma 3 1B PT 22.2%
  119. 119 DeepSeek-R1-Distill-Qwen-32B 21.3%
  120. 120 Qwen1.5-14B 21.1%
  121. 121 DeepSeek-R1-Distill-Qwen-14B 21.1%
  122. 122 Tiny Aya Global val 21.1%
  123. 123 Gemma 2 2B IT 20%
  124. 124 Qwen3.5-2B-Base val 19.5%
  125. 125 Tiny Aya Water val 19.3%
  126. 126 DeepSeek-R1-Distill-Qwen-7B 17.8%
  127. 127 Qwen3-0.6B 16.1%
  128. 128 Qwen3.5-0.8B val 15%
  129. 129 Gemma 3 1B 12.8%
  130. 130 LFM2-1.2B 12.5%
  131. 131 Mistral-7B-Instruct-v0.1 11.4%
  132. 132 Phi-4-reasoning 10.2%
  133. 133 LFM2.5-1.2B-Thinking 7.00%
  134. 134 LFM2.5-8B-A1B val 6.08%
  135. 135 SmolLM2 1.7B Instruct 5.78%
  136. 136 Qwen3.5-0.8B-Base val 5.13%
  137. 137 Gemma 3 270M 5.04%
  138. 138 Llama 3.2 1B 4.85%
  139. 139 Falcon3-3B-Instruct 3.87%
  140. 140 DeepSeek-R1-Distill-Qwen-1.5B 3.04%
  141. 141 SmolLM2-360M 2.97%
  142. 142 Phi-4-Reasoning-plus 2.32%
  143. 143 SmolLM2-135M 1.38%
  144. 144 SmolLM2-1.7B 0.77%
  145. 145 Phi-4-mini-reasoning 0.66%
  146. 146 Phi-2 -0.08%
  147. 147 LFM2.5-1.2B-Instruct -1.39%

Source: Official FoSent leaderboard

#EntryScore ±SettingDateSource
1 Gemini 2.5 Pro Jun 2025 91.6% ±1.9 zero-shot, val Jun 2025 Official leaderboard
2 speakleash/Bielik-11B-v3.1-Instruct 84.4% val Official leaderboard
3 Gemini 3 Pro 83.1% ±2 zero-shot, val Nov 2025 Official leaderboard
4 gemini/gemini-3-flash-preview#thinking 82.8% ±1.9 zero-shot, val Official leaderboard
5 Gemini 3.7 Flash 81.8% ±2.3 zero-shot, val Aug 2026 Official leaderboard
6 deepseek/deepseek-flash#no-thinking 81.7% ±2 val Official leaderboard
7 o3 81% ±2 zero-shot, val Apr 2025 Official leaderboard
8 gemini/gemini-3-flash-preview#no-thinking 79.6% ±2 zero-shot, val Official leaderboard
9 Qwen/Qwen3.6-27B-FP8 78.6% ±2.6 val Official leaderboard
10 NbAiLab/borealis2-31b-preview 78.5% ±3 val Official leaderboard
11 GPT-6.2 Astra 78% ±1.7 zero-shot, val Nov 2026 Official leaderboard
12 Claude Sonnet 4.6 77.8% ±1.7 zero-shot, val Feb 2026 Official leaderboard
13 GPT-4.1 77.8% ±2.3 zero-shot, val Apr 2025 Official leaderboard
14 GPT-6 Astra 77.7% ±1.9 zero-shot, val Sep 2026 Official leaderboard
15 GPT-5 mini 77.7% ±2 zero-shot, val Aug 2025 Official leaderboard
16 GPT-5.6 Terra 77.5% ±2.3 zero-shot, val Jul 2026 Official leaderboard
17 openai/gpt-5.4-mini-2026-03-17#low 76.4% ±2.7 zero-shot, val Official leaderboard
18 Gemma 4 31B IT 76.3% ±2.4 val Apr 2026 Official leaderboard
19 GPT-6 Sol 75.9% ±2 zero-shot, val Sep 2026 Official leaderboard
20 Gemini 3.1 Flash-Lite 75.3% ±1.9 zero-shot, val Mar 2026 Official leaderboard
21 speakleash/Bielik-11B-v3.0-Instruct 74.8% val Official leaderboard
22 claude-sonnet-4-5-20250929#no-thinking 74.6% ±3 zero-shot, val Official leaderboard
23 deepseek/deepseek-flash#low 74.6% ±2.2 zero-shot, val Official leaderboard
24 claude-sonnet-4-5-20250929#thinking 74.3% ±1.8 zero-shot, val Official leaderboard
25 openai/gpt-5.4-mini-2026-03-17#high 74% ±3.6 zero-shot, val Official leaderboard
26 Gemini 3.6 Flash 73.9% ±2 zero-shot, val Jul 2026 Official leaderboard
27 Grok-2 Dec 2024 73.8% ±2 zero-shot, val Dec 2024 Official leaderboard
28 openai/gpt-5.4-mini-2026-03-17#medium 73.5% ±2.6 zero-shot, val Official leaderboard
29 GPT-6.1 Sol 73.2% ±1.5 zero-shot, val Sep 2026 Official leaderboard
30 openai/gpt-5.4-mini-2026-03-17#none 73% ±3.3 zero-shot, val Official leaderboard
31 gpt-5-mini-2025-08-07#high 72.9% ±1.9 zero-shot, val Official leaderboard
32 Qwen/Qwen3.5-35B-A3B-FP8 72.9% ±2.8 val Official leaderboard
33 GLM-5.3-Flash 72.8% ±4.3 val Aug 2026 Official leaderboard
34 Gemma 4 12B 72.4% ±5.2 val Jun 2026 Official leaderboard
35 NbAiLab/borealis2-26b-a4b-preview 71.7% ±2.6 val Official leaderboard
36 Gemma 4 26B A4B 71.4% ±1.3 val Apr 2026 Official leaderboard
37 gemini/gemini-2.0-flash-001 71.4% ±1.8 zero-shot, val Official leaderboard
38 GPT-5.2 71.3% ±2 zero-shot, val Dec 2025 Official leaderboard
39 mistralai/Mistral-Large-3-675B-Instruct-2512 71% ±3.2 val Official leaderboard
40 claude-3-7-sonnet-20250219#thinking 71% ±1.9 zero-shot, val Official leaderboard
41 openai/gpt-5.4-nano-2026-03-17#medium 71% ±2.8 zero-shot, val Official leaderboard
42 GPT-6.2 Sol 70.9% ±2.4 zero-shot, val Nov 2026 Official leaderboard
43 Gemini 3.5 Flash-Lite 70.3% ±3 zero-shot, val Jul 2026 Official leaderboard
44 o4-mini 69.6% ±2.8 zero-shot, val Apr 2025 Official leaderboard
45 Apertus 1.5 8B 68.7% ±2.7 val Jul 2026 Official leaderboard
46 gemini/gemini-2.0-flash-lite-001 68.6% ±2.1 zero-shot, val Official leaderboard
47 gemini/gemini-2.5-flash#no-thinking 68.3% ±2.4 zero-shot, val Official leaderboard
48 openai/gpt-5.4-nano-2026-03-17#low 67.9% ±3.2 zero-shot, val Official leaderboard
49 GadflyII/GLM-4.7-Flash-NVFP4 67.8% ±3.4 val Official leaderboard
50 gpt-5-2025-08-07#high 67% ±2.7 zero-shot, val Official leaderboard
51 Claude 3.5 Haiku 66% ±2.3 zero-shot, val Oct 2024 Official leaderboard
52 danish-foundation-models/DFM-Mimir-v1.8 65.9% ±2 val Official leaderboard
53 NbAiLab/borealis-12b 65.5% ±4.4 val Official leaderboard
54 glm-4.7-flash 65.2% ±3.7 val Jan 2026 Official leaderboard
55 GPT-5 65.1% ±2.8 zero-shot, val Aug 2025 Official leaderboard
56 Qwen3.5-35B-A3B-Base 64.9% ±5.7 val Feb 2026 Official leaderboard
57 gpt-5-mini-2025-08-07#minimal 64.4% ±2.6 zero-shot, val Official leaderboard
58 Mistral Medium 3.5 64.2% ±2.8 val Apr 2026 Official leaderboard
59 mistralai/Ministral-3-14B-Instruct-2512 64.1% ±1.3 zero-shot, val Official leaderboard
60 Cogito v1 Preview 70B 63.9% ±1.6 Apr 2025 Official leaderboard
61 Claude Haiku 4.5 63.4% ±2 zero-shot, val Oct 2025 Official leaderboard
62 Qwen3.5-9B 63.3% ±4.4 val Feb 2026 Official leaderboard
63 ollama_chat/gemma3:27b-it-fp16 63.2% ±1.3 Official leaderboard
64 NbAiLab/borealis-open-12b 62.8% ±4.9 val Official leaderboard
65 Unbabel/Tower-Plus-9B 62.7% ±3.4 val Official leaderboard
66 GPT-6 Luna 62.5% ±2 zero-shot, val Sep 2026 Official leaderboard
67 gpt-5-2025-08-07#minimal 62.4% ±2.3 zero-shot, val Official leaderboard
68 claude-3-7-sonnet-20250219#no-thinking 62.3% ±1.5 zero-shot, val Official leaderboard
69 NbAiLab/borealis-27b 62.2% ±2.8 val Official leaderboard
70 Gemini 1.5 Flash Sep 2024 62.1% ±3.5 zero-shot, val Aug 2024 Official leaderboard
71 Qwen 3.8 27B 62.1% ±2.7 val Aug 2026 Official leaderboard
72 NbAiLab/borealis-27b-instruct-preview 62% ±2.5 val Official leaderboard
73 NbAiLab/borealis2-e4b-preview 61.9% ±3.7 val Official leaderboard
74 danish-foundation-models/DFM-Mimir-v1.5 61.8% ±2 zero-shot, val Official leaderboard
75 NbAiLab/borealis-12b-instruct-preview 61.8% ±3.2 val Official leaderboard
76 NbAiLab/borealis-open-27b 61.5% ±3.4 val Official leaderboard
77 Grok 3 61.3% ±1.8 zero-shot, val Apr 2025 Official leaderboard
78 Gemma 3 27B 61.2% ±1 zero-shot Mar 2025 Official leaderboard
79 gemini/gemini-2.5-flash-lite-preview-06-17#thinking 61.1% ±3 zero-shot, val Official leaderboard
80 Gemini 2.5 Flash-Lite Sep 2025 61.1% ±3 zero-shot, val Sep 2025 Official leaderboard
81 LumiOpen/Llama-Poro-2-70B-Instruct 60.5% ±0.9 Official leaderboard
82 Gemma 4 E4B 60.2% ±3.4 val Mar 2026 Official leaderboard
83 o3-mini 59.9% ±2.5 zero-shot, val Jan 2025 Official leaderboard
84 GPT-4.1 mini 59.6% ±2.6 zero-shot, val Apr 2025 Official leaderboard
85 GPT-5.6 Sol 59.5% ±2.1 zero-shot, val Jul 2026 Official leaderboard
86 deepcogito/cogito-v2-preview-llama-70B 58.8% ±1.2 Official leaderboard
87 meta-llama/Llama-3.1-405B-Instruct-FP8 58.7% ±1.1 Official leaderboard
88 Apertus 1.5 70B 58.7% ±1.9 val Jul 2026 Official leaderboard
89 ThatsGroes/gemma-2-27b-it-FP8-Dynamic 58.5% ±0.9 Official leaderboard
90 Llama 3.1-405B 58.1% ±2 Jul 2024 Official leaderboard
91 google/gemma-2-27b-it 57.8% ±1.1 Official leaderboard
92 openai/gpt-5.4-2026-03-05#none 57.1% ±2.9 zero-shot, val Official leaderboard
93 typesafe/jev-1.13.0 57% ±1.4 zero-shot Official leaderboard
94 Gemini 2.5 Flash Sep 2025 57% ±2.8 zero-shot, val Sep 2025 Official leaderboard
95 Qwen/Qwen3-235B-A22B-Instruct-2507-FP8 56.5% ±1.1 Official leaderboard
96 Gemma 2 27B 56.5% ±2.1 Jun 2024 Official leaderboard
97 zai-org/GLM-4.5-Air#no-thinking 56.4% ±1 Official leaderboard
98 gpt-oss-120b 56.4% ±3.2 val Aug 2025 Official leaderboard
99 mistralai/Mistral-Small-4-119B-2603 56.3% ±3.2 val Official leaderboard
100 LumiOpen/Llama-Poro-2-70B-SFT 56% ±1.4 Official leaderboard
101 GPT-5 nano 55.9% ±2.9 zero-shot, val Aug 2025 Official leaderboard
102 Qwen/Qwen3-235B-A22B-Thinking-2507-FP8 55.8% ±1.5 Official leaderboard
103 Ministral 3 8B Instruct 55.7% ±2.7 val Dec 2025 Official leaderboard
104 Gemma 3 12B PT 55.4% ±0.8 Mar 2025 Official leaderboard
105 NovaSky-AI/Sky-T1-32B-Preview 55.2% ±1.4 Official leaderboard
106 mistralai/Ministral-3-8B-Base-2512 54.8% ±4.9 val Official leaderboard
107 google/gemma-2-9b-it 54.6% ±1.1 Official leaderboard
108 Qwen3-235B-A22B 54.4% ±2 Apr 2025 Official leaderboard
109 Gemma 3 12B 54.1% ±1.4 Mar 2025 Official leaderboard
110 Unbabel/Tower-Plus-72B 54.1% ±3.3 val Official leaderboard
111 Qwen3.5-4B 53.9% ±6 val Feb 2026 Official leaderboard
112 Llama 4 Scout 53.3% ±1.1 Apr 2025 Official leaderboard
113 HPLT/NorOLMo-13B 53.3% ±2.4 val Official leaderboard
114 Grok 4.1 Fast 53.3% ±2.1 zero-shot, val Nov 2025 Official leaderboard
115 meta-llama/Llama-4-Scout-17B-16E 53.2% ±1.6 Official leaderboard
116 Gemma 3n E4B 53% ±2.1 Jun 2025 Official leaderboard
117 deepcogito/cogito-v1-preview-qwen-32B 52.8% ±1.2 Official leaderboard
118 Llama 3.1-70B 52.7% ±2.2 Jul 2024 Official leaderboard
119 Grok-3 mini 52.6% ±2.5 zero-shot, val Apr 2025 Official leaderboard
120 Granite 4.0 H Small 52.3% ±1.4 Oct 2025 Official leaderboard
121 Gemma 3 27B PT 52.2% ±2.9 Mar 2025 Official leaderboard
122 mistralai/Ministral-3-14B-Base-2512 52.1% ±6.4 val Official leaderboard
123 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8 51.9% ±4.3 val Official leaderboard
124 openai/gpt-5.4-nano-2026-03-17#none 51.6% ±2.1 zero-shot, val Official leaderboard
125 gemini/gemini-2.5-flash-lite-preview-06-17#no-thinking 51.2% ±2 zero-shot, val Official leaderboard
126 gemini/gemini-2.5-flash-lite#no-thinking 51.2% ±2 zero-shot, val Official leaderboard
127 Phi-4 51% ±1.4 Dec 2024 Official leaderboard
128 mistralai/Mistral-Small-24B-Instruct-2501 50.8% ±1.7 Official leaderboard
129 swiss-ai/Apertus-v1.1-4B 50.7% ±3.3 val Official leaderboard
130 deepcogito/cogito-v1-preview-qwen-14B 50.6% ±1.7 Official leaderboard
131 Qwen3-Next-80B-A3B-Instruct 50.4% ±0.6 Sep 2025 Official leaderboard
132 Nemotron-Cascade 2 30B-A3B 50.2% ±2.7 val Mar 2026 Official leaderboard
133 Gemma 3n E2B 50.2% ±1 Jun 2025 Official leaderboard
134 Qwen/Qwen3-32B#no-thinking 50.1% ±1.3 Official leaderboard
135 Apertus 70B 49.8% ±0.7 Sep 2025 Official leaderboard
136 bullerwins/Magistral-Small-2506-fp8 49.4% ±1.1 Official leaderboard
137 mistralai/Mistral-Nemo-Instruct-2407 48.8% ±1.6 Official leaderboard
138 meta-llama/Llama-3.1-8B-Instruct 48.2% ±1 Official leaderboard
139 Qwen3-Next-80B-A3B 47.7% ±1.9 thinking Sep 2025 Official leaderboard
140 QwQ-32B 47.7% ±1.3 Mar 2025 Official leaderboard
141 allenai/OLMo-2-0325-32B 47.6% ±1.1 Official leaderboard
142 mistralai/Mistral-Small-3.1-24B-Instruct-2503 47.3% ±1 Official leaderboard
143 Aya Expanse 32B 47.2% ±1.5 Oct 2024 Official leaderboard
144 danish-foundation-models/DFM-Mimir-v1.8-Mini 47.2% ±1.4 zero-shot, val Official leaderboard
145 mistralai/Mistral-Small-3.2-24B-Instruct-2506 47.1% ±1.6 Official leaderboard
146 152334H/miqu-1-70b-sf 46.8% ±2 Official leaderboard
147 Llama-3.1-Tulu-3-8B 46.5% ±1.1 Nov 2024 Official leaderboard
148 Gemma 3 4B 46.5% ±1.4 Mar 2025 Official leaderboard
149 allenai/Llama-3.1-Tulu-3.1-8B 46.4% ±0.9 Official leaderboard
150 allenai/Llama-3.1-Tulu-3-8B-DPO 46.4% ±1.1 Official leaderboard
151 GLM-4.5-Air 46.2% ±1.7 thinking Jul 2025 Official leaderboard
152 mistralai/Mistral-Small-3.1-24B-Base-2503 46.1% ±1.8 Official leaderboard
153 ollama_chat/gemma3:4b-it-q4_K_M 46% ±1.2 Official leaderboard
154 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 46% ±4.5 val Official leaderboard
155 mistralai/Magistral-Small-2509 45.8% ±1.3 Official leaderboard
156 utter-project/EuroLLM-22B-Instruct-2512 45.6% ±1.6 Official leaderboard
157 ollama_chat/phi4:14b-q4_K_M 45.6% ±4.7 val Official leaderboard
158 Qwen3.5-4B-Base 45.5% ±2.5 val Feb 2026 Official leaderboard
159 Qwen/Qwen3-30B-A3B#no-thinking 45.4% ±1.6 Official leaderboard
160 NbAiLab/borealis-4b-instruct-preview 45.4% ±1.9 Official leaderboard
161 Reka Flash 3 44.9% ±1.2 Mar 2025 Official leaderboard
162 Voxtral Small 44.8% ±1.3 Jul 2025 Official leaderboard
163 ollama_chat/gemma3:12b-it-fp16 44.8% ±4.4 val Official leaderboard
164 Gemma 2 9B 44.7% ±2.1 Jun 2024 Official leaderboard
165 Apertus 8B Instruct 44.5% ±1.4 Sep 2025 Official leaderboard
166 Reka Flash 3.1 44.4% ±1.6 May 2025 Official leaderboard
167 ollama_chat/hf.co/google/gemma-3-4b-it-qat-q4_0-gguf 44.3% ±1.3 Official leaderboard
168 Gemma 3 4B PT 44% ±1.8 Mar 2025 Official leaderboard
169 ollama_chat/gemma3:27b-it-q4_K_M 43.8% ±6.1 val Official leaderboard
170 domyn/Domyn-Small-v1.0 43.7% ±3.8 val Official leaderboard
171 Qwen/Qwen3-14B#no-thinking 43.5% ±1.9 Official leaderboard
172 OLMo 2 32B Instruct 43.5% ±1.7 Mar 2025 Official leaderboard
173 gpt-5-nano-2025-08-07#high 43.3% ±3.2 zero-shot, val Official leaderboard
174 deepseek-r1-0528-qwen3-8b 43.1% ±1.4 May 2025 Official leaderboard
175 openai/gpt-5.4-nano-2026-03-17#high 42.8% ±2.7 zero-shot, val Official leaderboard
176 Gemma 7B 42.4% ±1.9 Feb 2024 Official leaderboard
177 Ministral 3 14B 42.4% ±4.8 val Dec 2025 Official leaderboard
178 norallm/normistral-11b-thinking 42.1% ±3 Official leaderboard
179 GPT-5.6 Luna 42.1% ±1.6 zero-shot, val Jul 2026 Official leaderboard
180 mistralai/Ministral-3-8B-Reasoning-2512 41.8% ±1.8 Official leaderboard
181 danish-foundation-models/DFM-Mimir 41.3% ±3.9 val Official leaderboard
182 Qwen/Qwen3-8B#no-thinking 41.1% ±1.1 Official leaderboard
183 Qwen3.5-9B-Base 40.7% ±2.4 val Feb 2026 Official leaderboard
184 norallm/normistral-11b-long 40.3% ±2.8 Official leaderboard
185 Unbabel/Tower-Plus-2B 40.2% ±4.3 val Official leaderboard
186 baidu/ERNIE-4.5-21B-A3B-PT 40.1% ±1.6 Official leaderboard
187 deepcogito/cogito-v1-preview-llama-8B 40.1% ±1.9 Official leaderboard
188 GLM-4.1V-9B-Thinking 40% ±1.2 Jul 2025 Official leaderboard
189 Qwen3-8B 40% ±1.8 Apr 2025 Official leaderboard
190 NousResearch/DeepHermes-3-Llama-3-8B-Preview 39.8% ±1.4 Official leaderboard
191 swiss-ai/Apertus-v1.1-4B-Instruct 39.8% ±3.4 val Official leaderboard
192 ibm-granite/granite-4.0-h-tiny 39.6% ±1.6 Official leaderboard
193 mistralai/Ministral-3-3B-Base-2512 39.5% ±1.5 Official leaderboard
194 utter-project/EuroLLM-22B-Instruct-Preview 39.4% ±1.4 Official leaderboard
195 speakleash/Bielik-11B-v2.6-Instruct 39.3% ±1.4 Official leaderboard
196 ollama_chat/gemma3:12b-it-q4_K_M 39.3% ±3.7 val Official leaderboard
197 norallm/normistral-11b-warm 39% ±4.1 Official leaderboard
198 openGPT-X/Teuken-7B-instruct-v0.6 38.9% ±5.5 val Official leaderboard
199 TrustLLMeu/baseline-7-8b_1t-tokens_llama 38.6% ±3.6 Official leaderboard
200 Qwen3-14B 38.5% ±4.6 Apr 2025 Official leaderboard
201 speakleash/Bielik-11B-v2.3-Instruct 38.4% ±1.1 Official leaderboard
202 Gemma 4 E2B 37.9% ±3.8 val Mar 2026 Official leaderboard
203 AI-Sweden-Models/Llama-3-8B-instruct 37.7% ±2.9 val Official leaderboard
204 DeepSeek-R1-Distill-Llama-70B 37.2% ±1.9 Jan 2025 Official leaderboard
205 gpt-oss-20b 37.2% ±1 medium Aug 2025 Official leaderboard
206 NbAiLab/borealis-4b 37.1% ±3.4 val Official leaderboard
207 TrustLLMeu/baseline-7-8b_2-3t-tokens_llama 37% ±3.5 Official leaderboard
208 granite-4.0-micro 36.7% ±2.3 Oct 2025 Official leaderboard
209 nvidia-nemotron-nano-9b-v2 36.6% ±2.1 Aug 2025 Official leaderboard
210 utter-project/EuroLLM-22B-Preview 36.5% ±1.6 Official leaderboard
211 Qwen3.5-2B 36.5% ±4.9 val Feb 2026 Official leaderboard
212 meta-llama/Llama-2-70b-hf 36.3% ±3.9 Official leaderboard
213 Llama 3.1-8B 36.3% ±2.3 Jul 2024 Official leaderboard
214 ai21labs/Jamba-v0.1 36.1% ±2.4 Official leaderboard
215 amalia-llm/AMALIA-9B-0626-DPO 35.9% ±3.5 zero-shot, val Official leaderboard
216 NbAiLab/borealis-open-4b 35.9% ±3.5 val Official leaderboard
217 utter-project/EuroLLM-9B-Instruct 35.7% ±4.5 Official leaderboard
218 Qwen1.5-14B-Chat 35.6% ±1.1 Feb 2024 Official leaderboard
219 Qwen3-4B-Thinking-2507 35.4% ±1.9 Aug 2025 Official leaderboard
220 Ministral 3 3B Instruct 34.8% ±5.8 val Dec 2025 Official leaderboard
221 GPT-4.1 nano 34.7% ±1.9 zero-shot, val Apr 2025 Official leaderboard
222 bineric/lynx-instruct-30b#thinking 34.6% ±2.9 Official leaderboard
223 allenai/Olmo-3.1-32B-Instruct-DPO 34.6% ±1.7 Official leaderboard
224 utter-project/EuroLLM-9B 34% ±2.5 Official leaderboard
225 Falcon3-10B-Instruct 34% ±2.6 Dec 2024 Official leaderboard
226 amalia-llm/AMALIA-9B-0626-SFT 33.1% ±3.2 zero-shot, val Official leaderboard
227 speakleash/Bielik-Minitron-7B-v3.0-Instruct 33.1% val Official leaderboard
228 swiss-ai/Apertus-v1.1-1.5B 33.1% ±5.1 val Official leaderboard
229 allenai/Llama-3.1-Tulu-3-8B-SFT 32.9% ±2 Official leaderboard
230 tiiuae/falcon-11B 32.7% ±1.8 Official leaderboard
231 qwen3-4b-instruct-2507 32.5% ±1.5 Aug 2025 Official leaderboard
232 bineric/lynx-instruct-30b#no-thinking 32.5% ±2.4 Official leaderboard
233 Olmo 3.1 32B Instruct 32.4% ±1.9 Dec 2025 Official leaderboard
234 Command R7B 32.2% ±1.6 Dec 2024 Official leaderboard
235 senseable/WestLake-7B-v2 32% ±2.9 Official leaderboard
236 allenai/Olmo-3.1-32B-Instruct-SFT 32% ±1 Official leaderboard
237 mistralai/Voxtral-Mini-3B-2507 31.7% ±1.7 Official leaderboard
238 EXAONE 4.0 32B 31.6% ±3.9 val Jul 2025 Official leaderboard
239 Falcon3-7B-Instruct 31.5% ±1.6 Dec 2024 Official leaderboard
240 upstage/SOLAR-10.7B-v1.0 31.4% ±3.6 Official leaderboard
241 Granite-3.0-8B-Instruct 31.1% ±3.7 Oct 2024 Official leaderboard
242 meta-llama/Llama-3.2-3B-Instruct 31% ±1.6 Official leaderboard
243 Qwen/Qwen3-4B#no-thinking 31% ±1.2 Official leaderboard
244 Qwen3-4B 31% ±1.2 Apr 2025 Official leaderboard
245 01-ai/Yi-1.5-9B 30.8% ±1.7 Official leaderboard
246 NbAiLab/borealis-1b-instruct-preview 30.6% ±4.2 val Official leaderboard
247 mii-llm/nesso-4B 30.5% ±1.5 Official leaderboard
248 HuggingFaceTB/SmolLM3-3B 30.3% ±1.1 Official leaderboard
249 deepcogito/cogito-v1-preview-llama-3B 30% ±2 Official leaderboard
250 Aya Expanse 8B 29.7% ±2.5 Oct 2024 Official leaderboard
251 ibm-granite/granite-34b-code-base 29.6% ±1.6 Official leaderboard
252 Qwen3-1.7B 29.5% ±1.3 Apr 2025 Official leaderboard
253 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-Base-BF16 29.1% ±3.8 val Official leaderboard
254 Olmo 3.1 32B Think 29% ±4 val Dec 2025 Official leaderboard
255 AFM-4.5B 29% ±1.5 Jul 2025 Official leaderboard
256 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6 28.6% ±1.9 Official leaderboard
257 Yi-1.5-34B 28.5% ±1.8 May 2024 Official leaderboard
258 mistralai/Ministral-8B-Instruct-2410 28.1% ±1.5 Official leaderboard
259 allenai/OLMo-2-1124-13B-Instruct 27.9% ±1.5 Official leaderboard
260 tiiuae/Falcon3-10B-Base 27.7% ±2 Official leaderboard
261 Qwen/Qwen3-1.7B#no-thinking 27.6% ±1 Official leaderboard
262 Granite-3.0-8B-Base 27.6% ±2.4 Oct 2024 Official leaderboard
263 Ministral 3 3B Reasoning 27.4% ±1.4 Dec 2025 Official leaderboard
264 deahmed/Qwen3.5-2B-da-task 27.3% ±3.9 val Official leaderboard
265 OpenLLM-France/Luciole-23B-Instruct-1.1 27.3% ±4 val Official leaderboard
266 Mistral 7B v0.3 26.7% ±2.4 Sep 2023 Official leaderboard
267 NbAiLab/borealis-open-1b 26.7% ±3.4 val Official leaderboard
268 speakleash/Bielik-4.5B-v3.0-Instruct 26.6% ±1.9 Official leaderboard
269 utter-project/EuroLLM-9B-Instruct-2512 26.5% ±2.9 val Official leaderboard
270 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16#no-thinking 26.3% ±2.2 Official leaderboard
271 Aleph-Alpha/Kolibri-1 26.2% ±5.4 val Official leaderboard
272 HuggingFaceTB/SmolLM3-3B#no-thinking 25.8% ±1.8 Official leaderboard
273 allenai/OLMo-2-1124-13B 25.8% ±2.1 Official leaderboard
274 Mistral 7B v0.1 25.5% ±3.1 Sep 2023 Official leaderboard
275 NbAiLab/borealis2-e2b-preview 25.4% ±3.1 zero-shot, val Official leaderboard
276 ibm-granite/granite-20b-code-base 25.3% ±1.6 Official leaderboard
277 ibm-granite/granite-34b-code-instruct 25.2% ±1.4 Official leaderboard
278 TildeAI/TildeOpen-30b 24.6% ±3.7 val Official leaderboard
279 ibm-granite/granite-4.0-h-micro 24.5% ±2.8 Official leaderboard
280 syvai/claes-v1 24.4% ±1.4 zero-shot Official leaderboard
281 allenai/OLMo-2-1124-7B-Instruct 24.4% ±1.9 Official leaderboard
282 LumiOpen/Viking-33B 24.1% ±4.9 Official leaderboard
283 arcee-ai/AFM-4.5B-Base 24% ±2 Official leaderboard
284 CohereForAI/aya-23-35B 24% ±2.1 Official leaderboard
285 Gemma 2B 23.9% ±2.6 Feb 2024 Official leaderboard
286 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660 23.8% ±1.5 Official leaderboard
287 Mistral-7B-Instruct-v0.2 23.8% ±1.8 Dec 2023 Official leaderboard
288 speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551 23.6% ±1.1 zero-shot Official leaderboard
289 ibm-granite/granite-8b-code-base-4k 23.4% ±2.3 Official leaderboard
290 tiiuae/Falcon3-7B-Base 23.1% ±3.5 Official leaderboard
291 CohereLabs/tiny-aya-base 22.8% ±3.8 val Official leaderboard
292 Gemma 3 1B PT 22.2% ±3.1 Mar 2025 Official leaderboard
293 bineric/lynx-instruct-30b 22.2% ±4.2 val Official leaderboard
294 NbAiLab/borealis-1b 22.1% ±5.3 val Official leaderboard
295 TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894 21.9% ±3.5 Official leaderboard
296 syvai/claes-preview 21.8% ±1 zero-shot Official leaderboard
297 google/gemma-2-2b 21.7% ±2.6 Official leaderboard
298 DeepSeek-R1-Distill-Qwen-32B 21.3% ±3.5 Jan 2025 Official leaderboard
299 OpenLLM-France/Luciole-8B-Instruct-1.1 21.2% ±5.3 val Official leaderboard
300 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f 21.1% ±1.9 Official leaderboard
301 Qwen1.5-14B 21.1% ±0.8 Feb 2024 Official leaderboard
302 DeepSeek-R1-Distill-Qwen-14B 21.1% ±2 Jan 2025 Official leaderboard
303 Tiny Aya Global 21.1% ±5.4 val Feb 2026 Official leaderboard
304 Qwen/Qwen2.5-1.5B-Instruct 20.8% ±1.7 Official leaderboard
305 allenai/Olmo-3-1125-32B 20.8% ±2.8 Official leaderboard
306 meta-llama/Llama-2-70b-chat-hf 20.3% ±3.7 Official leaderboard
307 swiss-ai/Apertus-v1.1-1.5B-Instruct 20.2% ±5 val Official leaderboard
308 Gemma 2 2B IT 20% ±1.8 Jul 2024 Official leaderboard
309 Qwen3.5-2B-Base 19.5% ±2.3 val Feb 2026 Official leaderboard
310 Tiny Aya Water 19.3% ±5.6 val Feb 2026 Official leaderboard
311 allenai/OLMo-2-1124-7B 19.1% ±2 Official leaderboard
312 deepseek-ai/DeepSeek-R1-Distill-Llama-8B 19.1% ±1.7 Official leaderboard
313 nvidia/NVIDIA-Nemotron-3-Nano-4B-FP8 18.9% ±5.2 val Official leaderboard
314 meta-llama/Llama-2-13b-chat-hf 18.9% ±4.3 Official leaderboard
315 allenai/Olmo-3-1025-7B 18.8% ±2.4 Official leaderboard
316 allenai/Olmo-3-7B-Think 18.8% ±1.7 Official leaderboard
317 TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49 18.7% ±2.4 Official leaderboard
318 AI-Sweden-Models/gpt-sw3-6.7b-v2-translator 18.5% ±5.2 val Official leaderboard
319 nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16 18.1% ±1.1 Official leaderboard
320 LiquidAI/LFM2-2.6B 17.9% ±1.6 Official leaderboard
321 infly/OpenCoder-8B-Instruct 17.9% ±1.7 Official leaderboard
322 BSC-LT/ALIA-40b 17.8% ±2.6 Official leaderboard
323 allenai/Olmo-3-7B-Instruct-SFT 17.8% ±2.6 Official leaderboard
324 DeepSeek-R1-Distill-Qwen-7B 17.8% ±1.9 Jan 2025 Official leaderboard
325 utter-project/EuroLLM-22B-2512 17.7% ±5.1 Official leaderboard
326 ibm-granite/granite-8b-code-instruct-4k 17.6% ±2.4 Official leaderboard
327 swiss-ai/Apertus-8B-2509 17.5% ±2.6 Official leaderboard
328 allenai/Olmo-3-7B-Instruct-DPO 16.9% ±2.1 Official leaderboard
329 infly/OpenCoder-8B-Base 16.8% ±2.3 Official leaderboard
330 Qwen/Qwen2.5-Math-7B 16.5% ±1.9 Official leaderboard
331 MaLA-LM/emma-500-llama2-7b 16.3% ±4.3 Official leaderboard
332 Qwen3-0.6B 16.1% ±1.7 Apr 2025 Official leaderboard
333 allenai/Olmo-3-7B-Instruct 15.7% ±1.3 Official leaderboard
334 TildeAI/TildeOpen-30b#slow-tokenizer 15.5% ±2.5 Official leaderboard
335 NorwAI/NorwAI-Mixtral-8x7B-one-third-epoch 15.2% ±3.3 val Official leaderboard
336 openeurollm/oellm-9b-128k-theta32m-v3 15.2% ±5.3 val Official leaderboard
337 convaiinnovations/laya-multilingual 15.1% ±1 zero-shot Official leaderboard
338 Qwen3.5-0.8B 15% ±5.3 val Feb 2026 Official leaderboard
339 OpenLLM-France/Luciole-8B-Base 14.9% ±4.8 val Official leaderboard
340 BSC-LT/salamandra-7b-instruct 14.2% ±2 Official leaderboard
341 ibm-granite/granite-20b-code-instruct 14.2% ±0.2 Official leaderboard
342 speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c 13.5% ±0.9 zero-shot Official leaderboard
343 mideind/IceBERT-xlmr-ic3 13.4% ±2.5 Official leaderboard
344 openGPT-X/Teuken-7B-instruct-commercial-v0.4 13.2% ±3.9 Official leaderboard
345 norallm/normistral-7b-scratch 12.9% ±6.3 val Official leaderboard
346 AI-Sweden-Models/ModernBERT-large 12.8% ±1.7 Official leaderboard
347 Gemma 3 1B 12.8% ±2 Mar 2025 Official leaderboard
348 LFM2-1.2B 12.5% ±2.1 Jul 2025 Official leaderboard
349 common-pile/comma-v0.1-2t 12.5% ±2 Official leaderboard
350 google/gemma-7b-it 12.2% ±2.4 Official leaderboard
351 LumiOpen/Viking-13B 12% ±5 Official leaderboard
352 intfloat/multilingual-e5-large 11.9% ±2.9 Official leaderboard
353 DiscoResearch/DiscoLM-70b 11.9% ±3.7 Official leaderboard
354 OpenLLM-France/Luciole-1B-Instruct-1.1 11.5% ±4.3 val Official leaderboard
355 Mistral-7B-Instruct-v0.1 11.4% ±3.5 Sep 2023 Official leaderboard
356 yulan-team/YuLan-Mini-Instruct 11.3% ±1.8 Official leaderboard
357 EuroBERT/EuroBERT-210m 10.8% ±1.6 Official leaderboard
358 intfloat/multilingual-e5-base 10.8% ±2.8 Official leaderboard
359 Qwen/Qwen3-0.6B#no-thinking 10.7% ±2.2 Official leaderboard
360 HPLT/hplt2c_dan_checkpoints 10.7% ±2 Official leaderboard
361 jhu-clsp/mmBERT-base 10.5% ±2.4 Official leaderboard
362 arcee-ai/Trinity-Nano-Preview 10.5% ±2.4 Official leaderboard
363 occiglot/occiglot-7b-eu5 10.4% ±3.9 Official leaderboard
364 timpal0l/Mistral-7B-v0.1-flashback-v2 10.3% ±2.4 Official leaderboard
365 arcee-ai/Trinity-Mini-Base 10.3% ±2.3 Official leaderboard
366 meta-llama/Llama-2-13b-hf 10.2% ±4.4 Official leaderboard
367 utter-project/EuroVLM-1.7B-Preview 10.2% ±1.9 Official leaderboard
368 Phi-4-reasoning 10.2% ±1.7 Apr 2025 Official leaderboard
369 gpt-5-nano-2025-08-07#minimal 9.79% ±2.7 zero-shot, val Official leaderboard
370 meta-llama/Llama-2-7b-chat-hf 9.66% ±2.5 Official leaderboard
371 LiquidAI/LFM2-700M 9.54% ±2 Official leaderboard
372 ltg/norbert3-base 9.47% ±2.2 Official leaderboard
373 tencent/Hunyuan-7B-Instruct 9.36% ±2.1 Official leaderboard
374 LiquidAI/LFM2.5-8B-A1B-Base 9.03% ±5.7 val Official leaderboard
375 EuroBERT/EuroBERT-610m 8.54% ±0.6 Official leaderboard
376 AI-Sweden-Models/gpt-sw3-20b 8.40% ±1.7 Official leaderboard
377 ibm-granite/granite-3.0-3b-a800m-instruct 7.94% ±1.5 Official leaderboard
378 NbAiLab/borealis-270m-instruct-preview 7.85% ±4.1 val Official leaderboard
379 tencent/Hunyuan-1.8B-Instruct 7.84% ±1.9 Official leaderboard
380 marksverdhei/modern-norbert3-large 7.66% ±1.4 Official leaderboard
381 NbAiLab/borealis-open-270m 7.21% ±2.4 val Official leaderboard
382 HuggingFaceTB/SmolLM3-3B-Base 7.20% ±1.5 Official leaderboard
383 swiss-ai/Apertus-v1.1-0.5B-Instruct 7.20% ±3.7 val Official leaderboard
384 utter-project/EuroLLM-1.7B-Instruct 7.10% ±4.1 val Official leaderboard
385 TurkuNLP/finnish-modernbert-large-seq-len-1024-117300-annealed 7.07% ±1.4 Official leaderboard
386 LFM2.5-1.2B-Thinking 7.00% ±2.5 Jan 2026 Official leaderboard
387 swiss-ai/Apertus-v1.1-0.5B 6.91% ±4.1 val Official leaderboard
388 ltg/norbert3-large 6.56% ±2.1 Official leaderboard
389 KBLab/roberta-base-swedish-cased 6.28% ±2 Official leaderboard
390 jhu-clsp/mmBERT-small 6.25% ±2.5 Official leaderboard
391 HPLT/hplt2c_isl_checkpoints 6.11% ±1.4 Official leaderboard
392 LFM2.5-8B-A1B 6.08% ±7 val May 2026 Official leaderboard
393 TurkuNLP/finnish-modernbert-tiny 5.99% ±1.4 Official leaderboard
394 common-pile/comma-v0.1-1t 5.95% ±2.9 Official leaderboard
395 SmolLM2 1.7B Instruct 5.78% ±1.9 Oct 2024 Official leaderboard
396 NbAiLab/nb-roberta-base-ncc 5.54% ±3.1 Official leaderboard
397 HPLT/hplt2c_swe_checkpoints 5.53% ±1.8 Official leaderboard
398 ibm-granite/granite-3.0-3b-a800m-base 5.38% ±1.9 Official leaderboard
399 mideind/IceBERT-ic3 5.33% ±2.8 Official leaderboard
400 fresh-xlm-roberta-base 5.29% ±1.8 Official leaderboard
401 google-bert/bert-base-multilingual-uncased 5.27% ±2.7 val Official leaderboard
402 vesteinn/DanskBERT 5.25% ±1.5 Official leaderboard
403 AI-Sweden-Models/gpt-sw3-126m 5.21% ±1.4 Official leaderboard
404 Qwen3.5-0.8B-Base 5.13% ±2.9 val Feb 2026 Official leaderboard
405 Gemma 3 270M 5.04% ±1.5 Aug 2025 Official leaderboard
406 neurocode/IsRoBERTa 5.03% ±2.6 Official leaderboard
407 Llama 3.2 1B 4.85% ±1.5 Sep 2024 Official leaderboard
408 occiglot/occiglot-7b-eu5-instruct 4.76% ±2.7 Official leaderboard
409 speakleash/Bielik-1.5B-v3.0-Instruct 4.70% ±1.5 Official leaderboard
410 answerdotai/ModernBERT-large 4.69% ±1.1 Official leaderboard
411 meta-llama/Llama-3.2-1B-Instruct 4.68% ±1.6 Official leaderboard
412 NbAiLab/nb-bert-base 4.54% ±1.4 Official leaderboard
413 Maltehb/aelaectra-danish-electra-small-cased 4.53% ±1.3 Official leaderboard
414 TurkuNLP/finnish-modernbert-base 4.50% ±2 Official leaderboard
415 mideind/IceBERT-large 4.46% ±3.2 Official leaderboard
416 facebook/MobileLLM-R1-360M-base 4.33% ±1.8 Official leaderboard
417 answerdotai/ModernBERT-base 4.23% ±2.5 Official leaderboard
418 vesteinn/FoBERT 4.15% ±2.3 Official leaderboard
419 TurkuNLP/finnish-modernbert-large 4.02% ±1.9 Official leaderboard
420 kyutai/helium-1-2b 3.97% ±0.9 Official leaderboard
421 AI-Sweden-Models/roberta-large-1160k 3.89% ±2.1 Official leaderboard
422 Falcon3-3B-Instruct 3.87% ±2 Dec 2024 Official leaderboard
423 fastino/GLiNER2.5-Decide 3.83% ±1.1 zero-shot Official leaderboard
424 NbAiLab/nb-bert-large 3.74% ±2.2 Official leaderboard
425 openbmb/MiniCPM5-1B 3.73% ±5.7 val Official leaderboard
426 NbAiLab/nb-roberta-base-ncc-plus-scandi-1e4 3.71% ±2.9 Official leaderboard
427 arcee-ai/Trinity-Nano-Base 3.63% ±1.8 Official leaderboard
428 OpenLLM-France/Luciole-23B-Base 3.58% ±6.9 val Official leaderboard
429 baidu/ERNIE-4.5-0.3B-PT 3.55% ±1.8 Official leaderboard
430 speakleash/Bielik-11B-v3-Base-20250730 3.46% ±3.4 Official leaderboard
431 vesteinn/XLMR-ENIS 3.39% ±1.4 Official leaderboard
432 mideind/IceBERT-mC4-is 3.37% ±2.2 Official leaderboard
433 OmniJev/OneJev1.6-0.8B 3.33% ±0.6 zero-shot Official leaderboard
434 ltg/norbert3-small 3.29% ±1.6 Official leaderboard
435 AI-Sweden-Models/bert-large-nordic-pile-1M-steps 3.24% ±1.4 Official leaderboard
436 norallm/norbloom-7b-scratch 3.24% ±4 val Official leaderboard
437 KBLab/bert-base-swedish-cased 3.18% ±2.1 Official leaderboard
438 NbAiLab/notram-bert-norwegian-cased-080321 3.06% ±1.8 Official leaderboard
439 DeepSeek-R1-Distill-Qwen-1.5B 3.04% ±1.4 Jan 2025 Official leaderboard
440 ibm-granite/granite-3.0-1b-a400m-base 2.99% ±1.6 Official leaderboard
441 nvidia/NVIDIA-Nemotron-Nano-9B-v2#no-thinking 2.99% ±1.2 Official leaderboard
442 SmolLM2-360M 2.97% ±1.7 Oct 2024 Official leaderboard
443 HPLT/gpt-13b-nordic-prerelease 2.91% ±2 Official leaderboard
444 NbAiLab/nb-roberta-base-ncc-plus-scandi-2e4 2.82% ±4.1 Official leaderboard
445 jaredpalmer/kev-0.5b 2.78% ±0.9 zero-shot Official leaderboard
446 FacebookAI/xlm-roberta-large 2.72% ±1.4 Official leaderboard
447 vesteinn/ScandiBERT 2.69% ±1.4 Official leaderboard
448 KBLab/bert-base-swedish-cased-alpha 2.63% ±1.6 Official leaderboard
449 microsoft/mdeberta-v3-base 2.58% ±2 Official leaderboard
450 HPLT/hplt2c_nob_checkpoints 2.53% ±2.1 Official leaderboard
451 DDSC/roberta-base-danish 2.36% ±1.9 Official leaderboard
452 Phi-4-Reasoning-plus 2.32% ±1.6 Apr 2025 Official leaderboard
453 danish-foundation-models/encoder-small-v1 2.31% ±1.9 Official leaderboard
454 google/gemma-2b-it 2.29% ±1.5 Official leaderboard
455 HuggingFaceTB/SmolLM2-360M-Instruct 2.29% ±2.1 Official leaderboard
456 KBLab/bert-base-swedish-cased-new 2.28% ±2.3 Official leaderboard
457 Maltehb/danish-bert-botxo 2.28% ±2.4 Official leaderboard
458 PleIAs/Pleias-1.2b-Preview 2.27% ±1.7 Official leaderboard
459 mideind/IceBERT-igc 2.17% ±1.6 Official leaderboard
460 AI-Sweden-Models/gpt-sw3-356m-instruct 2.13% ±2.6 Official leaderboard
461 KBLab/sentence-bert-swedish-cased 2.06% ±2.2 Official leaderboard
462 Muse-research/Muse-3B 1.99% ±6.2 val Official leaderboard
463 NbAiLab/nb-roberta-base-scandi 1.97% ±2.7 Official leaderboard
464 utter-project/EuroMoE-2.6B-A0.6B-Preview 1.96% ±1 Official leaderboard
465 DGurgurov/xlm-r_dan-latn 1.94% ±1.6 Official leaderboard
466 NbAiLab/nb-roberta-base-scandi-1e4 1.82% ±2.5 Official leaderboard
467 DDSC/roberta-base-scandinavian 1.73% ±2.2 Official leaderboard
468 EuropeanParliament/EUBERT 1.72% ±1.7 Official leaderboard
469 LiquidAI/LFM2.5-1.2B-Base 1.63% ±1.5 Official leaderboard
470 Geotrend/distilbert-base-25lang-cased 1.59% ±1.5 Official leaderboard
471 AI-Sweden-Models/gpt-sw3-40b 1.54% ±2.8 val Official leaderboard
472 facebook/MobileLLM-R1-140M-base 1.40% ±1.8 Official leaderboard
473 Qwen/Qwen2.5-Math-1.5B 1.39% ±0.8 Official leaderboard
474 SmolLM2-135M 1.38% ±1.9 Oct 2024 Official leaderboard
475 AIDC-AI/Marco-Mini-Global-Base 1.37% ±1.9 val Official leaderboard
476 OpenLLM-France/Luciole-1B-Base 1.24% ±4.5 val Official leaderboard
477 infly/OpenCoder-1.5B-Base 1.23% ±1.6 Official leaderboard
478 mideind/IceBERT 1.12% ±1.9 Official leaderboard
479 LiquidAI/LFM2-350M 1.05% ±2 Official leaderboard
480 Qwen/Qwen1.5-0.5B 1.05% ±2.5 Official leaderboard
481 AI-Sweden-Models/gpt-sw3-126m-instruct 1.04% ±1.1 Official leaderboard
482 KBLab/megatron-bert-base-swedish-cased-125k 0.84% ±1.9 Official leaderboard
483 Qwen/Qwen1.5-0.5B-Chat 0.81% ±1.8 Official leaderboard
484 convaiinnovations/laya 0.78% ±1.1 zero-shot Official leaderboard
485 SmolLM2-1.7B 0.77% ±2.1 Oct 2024 Official leaderboard
486 HuggingFaceTB/SmolLM2-135M-Instruct 0.69% ±2.6 Official leaderboard
487 PleIAs/Baguettotron@refs/pr/6#no-thinking 0.68% ±1.7 Official leaderboard
488 Phi-4-mini-reasoning 0.66% ±2.2 Apr 2025 Official leaderboard
489 vesteinn/ScandiBERT-no-faroese 0.64% ±1.4 Official leaderboard
490 ibm-granite/granite-3b-code-instruct-2k 0.44% ±0.4 Official leaderboard
491 ibm-granite/granite-3b-code-base-2k 0.34% ±0.3 Official leaderboard
492 marksverdhei/modern-norbert3-small 0.34% ±1.8 Official leaderboard
493 PleIAs/Pleias-350m-Preview 0.22% ±1.4 Official leaderboard
494 KBLab/megatron-bert-large-swedish-cased-165k 0.21% ±2 Official leaderboard
495 CohereLabs/aya-23-8B 0.19% ±6.4 val Official leaderboard
496 infly/OpenCoder-1.5B-Instruct 0.19% ±0.2 Official leaderboard
497 ltg/norbert3-xs 0.19% ±2.6 Official leaderboard
498 NbAiLab/nb-bert-ncc-male2female 0.18% ±1.6 Official leaderboard
499 AI-Sweden-Models/gpt-sw3-356m 0.13% ±1.7 Official leaderboard
500 facebook/MobileLLM-R1-140M 0.06% ±1.9 Official leaderboard
501 Maltehb/aelaectra-danish-electra-small-uncased 0.05% ±1.4 Official leaderboard
502 marksverdhei/modern-norbert3-base 0.04% ±2 Official leaderboard
503 AI-Sweden-Models/gpt-sw3-1.3b-instruct 0% val Official leaderboard
504 AI-Sweden-Models/gpt-sw3-6.7b-v2-instruct 0% val Official leaderboard
505 neuralmagic/Sparse-Llama-3.1-8B-2of4 0% Official leaderboard
506 NorwAI/NorwAI-Magistral-24B-reasoning 0% val Official leaderboard
507 openGPT-X/Teuken-7B-base-v0.6 0% val Official leaderboard
508 NbAiLab/roberta_jan_128_ncc -0.05% ±1.7 Official leaderboard
509 openai-community/gpt2-medium -0.08% ±1.2 Official leaderboard
510 Phi-2 -0.08% ±0.1 Dec 2023 Official leaderboard
511 NbAiLab/borealis-270m -0.09% ±3.6 val Official leaderboard
512 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview -0.10% ±1.2 Official leaderboard
513 ibm-granite/granite-3.0-1b-a400m-instruct -0.29% ±1.3 Official leaderboard
514 facebook/MobileLLM-R1-950M-base -0.33% ±0.4 Official leaderboard
515 KBLab/megatron-bert-base-swedish-cased-600k -0.47% ±1.3 Official leaderboard
516 marksverdhei/modern-norbert3-xs -0.58% ±1.4 Official leaderboard
517 openai-community/gpt2 -0.61% ±0.6 Official leaderboard
518 Shekswess/trlm-135m -0.68% ±1.7 Official leaderboard
519 openai-community/gpt2-large -0.81% ±1.1 Official leaderboard
520 facebook/MobileLLM-R1-360M -0.86% ±1.9 Official leaderboard
521 fresh-electra-small -1.03% ±1.4 Official leaderboard
522 inclusionAI/Ling-mini-2.0 -1.05% ±4 val Official leaderboard
523 LFM2.5-1.2B-Instruct -1.39% ±1.9 Jan 2026 Official leaderboard
524 AI-Sweden-Models/gpt-sw3-6.7b -1.42% ±1.4 val Official leaderboard
525 AI-Sweden-Models/gpt-sw3-6.7b-v2 -1.42% ±1.4 val Official leaderboard
526 ltg/norbert2 -1.92% ±2.2 Official leaderboard
527 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct -2.11% ±2 Official leaderboard
528 AI-Sweden-Models/gpt-sw3-1.3b -2.22% ±3.7 val Official leaderboard
529 ltg/norbert -2.30% ±1.2 Official leaderboard
530 distilbert/distilbert-base-multilingual-cased -2.34% ±1.7 Official leaderboard
531 HPLT/hplt2c_nno_checkpoints -2.70% ±2.7 Official leaderboard
532 utter-project/EuroLLM-1.7B -2.72% ±1.6 val Official leaderboard
533 BSC-LT/salamandra-7b -3.26% ±1.5 Official leaderboard
534 google-bert/bert-base-multilingual-cased -3.57% ±2.8 val Official leaderboard
535 tiiuae/Falcon3-3B-Base -3.57% ±1.2 Official leaderboard
536 PleIAs/Monad -4.18% ±1.4 Official leaderboard
537 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base -4.31% ±2.2 Official leaderboard
538 inclusionAI/Ling-mini-base-2.0 -5.17% ±5.2 val Official leaderboard
539 norallm/normistral-11b-translate -5.80% ±2.4 val Official leaderboard
540 facebook/MobileLLM-R1-950M -7.13% ±1 Official leaderboard
#ModelScore ±SettingDateSource
1 Gemini 2.5 Pro Jun 2025 91.6% ±1.9 zero-shot, val Jun 2025 Official leaderboard
2 Gemini 3 Pro 83.1% ±2 zero-shot, val Nov 2025 Official leaderboard
3 Gemini 3.7 Flash 81.8% ±2.3 zero-shot, val Aug 2026 Official leaderboard
4 o3 81% ±2 zero-shot, val Apr 2025 Official leaderboard
5 GPT-6.2 Astra 78% ±1.7 zero-shot, val Nov 2026 Official leaderboard
6 Claude Sonnet 4.6 77.8% ±1.7 zero-shot, val Feb 2026 Official leaderboard
7 GPT-4.1 77.8% ±2.3 zero-shot, val Apr 2025 Official leaderboard
8 GPT-6 Astra 77.7% ±1.9 zero-shot, val Sep 2026 Official leaderboard
9 GPT-5 mini 77.7% ±2 zero-shot, val Aug 2025 Official leaderboard
10 GPT-5.6 Terra 77.5% ±2.3 zero-shot, val Jul 2026 Official leaderboard
11 Gemma 4 31B IT 76.3% ±2.4 val Apr 2026 Official leaderboard
12 GPT-6 Sol 75.9% ±2 zero-shot, val Sep 2026 Official leaderboard
13 Gemini 3.1 Flash-Lite 75.3% ±1.9 zero-shot, val Mar 2026 Official leaderboard
14 Gemini 3.6 Flash 73.9% ±2 zero-shot, val Jul 2026 Official leaderboard
15 Grok-2 Dec 2024 73.8% ±2 zero-shot, val Dec 2024 Official leaderboard
16 GPT-6.1 Sol 73.2% ±1.5 zero-shot, val Sep 2026 Official leaderboard
17 GLM-5.3-Flash 72.8% ±4.3 val Aug 2026 Official leaderboard
18 Gemma 4 12B 72.4% ±5.2 val Jun 2026 Official leaderboard
19 Gemma 4 26B A4B 71.4% ±1.3 val Apr 2026 Official leaderboard
20 GPT-5.2 71.3% ±2 zero-shot, val Dec 2025 Official leaderboard
21 GPT-6.2 Sol 70.9% ±2.4 zero-shot, val Nov 2026 Official leaderboard
22 Gemini 3.5 Flash-Lite 70.3% ±3 zero-shot, val Jul 2026 Official leaderboard
23 o4-mini 69.6% ±2.8 zero-shot, val Apr 2025 Official leaderboard
24 Apertus 1.5 8B 68.7% ±2.7 val Jul 2026 Official leaderboard
25 Claude 3.5 Haiku 66% ±2.3 zero-shot, val Oct 2024 Official leaderboard
26 glm-4.7-flash 65.2% ±3.7 val Jan 2026 Official leaderboard
27 GPT-5 65.1% ±2.8 zero-shot, val Aug 2025 Official leaderboard
28 Qwen3.5-35B-A3B-Base 64.9% ±5.7 val Feb 2026 Official leaderboard
29 Mistral Medium 3.5 64.2% ±2.8 val Apr 2026 Official leaderboard
30 Cogito v1 Preview 70B 63.9% ±1.6 Apr 2025 Official leaderboard
31 Claude Haiku 4.5 63.4% ±2 zero-shot, val Oct 2025 Official leaderboard
32 Qwen3.5-9B 63.3% ±4.4 val Feb 2026 Official leaderboard
33 GPT-6 Luna 62.5% ±2 zero-shot, val Sep 2026 Official leaderboard
34 Gemini 1.5 Flash Sep 2024 62.1% ±3.5 zero-shot, val Aug 2024 Official leaderboard
35 Qwen 3.8 27B 62.1% ±2.7 val Aug 2026 Official leaderboard
36 Grok 3 61.3% ±1.8 zero-shot, val Apr 2025 Official leaderboard
37 Gemma 3 27B 61.2% ±1 zero-shot Mar 2025 Official leaderboard
38 Gemini 2.5 Flash-Lite Sep 2025 61.1% ±3 zero-shot, val Sep 2025 Official leaderboard
39 Gemma 4 E4B 60.2% ±3.4 val Mar 2026 Official leaderboard
40 o3-mini 59.9% ±2.5 zero-shot, val Jan 2025 Official leaderboard
41 GPT-4.1 mini 59.6% ±2.6 zero-shot, val Apr 2025 Official leaderboard
42 GPT-5.6 Sol 59.5% ±2.1 zero-shot, val Jul 2026 Official leaderboard
43 Apertus 1.5 70B 58.7% ±1.9 val Jul 2026 Official leaderboard
44 Llama 3.1-405B 58.1% ±2 Jul 2024 Official leaderboard
45 Gemini 2.5 Flash Sep 2025 57% ±2.8 zero-shot, val Sep 2025 Official leaderboard
46 Gemma 2 27B 56.5% ±2.1 Jun 2024 Official leaderboard
47 gpt-oss-120b 56.4% ±3.2 val Aug 2025 Official leaderboard
48 GPT-5 nano 55.9% ±2.9 zero-shot, val Aug 2025 Official leaderboard
49 Ministral 3 8B Instruct 55.7% ±2.7 val Dec 2025 Official leaderboard
50 Gemma 3 12B PT 55.4% ±0.8 Mar 2025 Official leaderboard
51 Qwen3-235B-A22B 54.4% ±2 Apr 2025 Official leaderboard
52 Gemma 3 12B 54.1% ±1.4 Mar 2025 Official leaderboard
53 Qwen3.5-4B 53.9% ±6 val Feb 2026 Official leaderboard
54 Llama 4 Scout 53.3% ±1.1 Apr 2025 Official leaderboard
55 Grok 4.1 Fast 53.3% ±2.1 zero-shot, val Nov 2025 Official leaderboard
56 Gemma 3n E4B 53% ±2.1 Jun 2025 Official leaderboard
57 Llama 3.1-70B 52.7% ±2.2 Jul 2024 Official leaderboard
58 Grok-3 mini 52.6% ±2.5 zero-shot, val Apr 2025 Official leaderboard
59 Granite 4.0 H Small 52.3% ±1.4 Oct 2025 Official leaderboard
60 Gemma 3 27B PT 52.2% ±2.9 Mar 2025 Official leaderboard
61 Phi-4 51% ±1.4 Dec 2024 Official leaderboard
62 Qwen3-Next-80B-A3B-Instruct 50.4% ±0.6 Sep 2025 Official leaderboard
63 Nemotron-Cascade 2 30B-A3B 50.2% ±2.7 val Mar 2026 Official leaderboard
64 Gemma 3n E2B 50.2% ±1 Jun 2025 Official leaderboard
65 Apertus 70B 49.8% ±0.7 Sep 2025 Official leaderboard
66 Qwen3-Next-80B-A3B 47.7% ±1.9 thinking Sep 2025 Official leaderboard
67 QwQ-32B 47.7% ±1.3 Mar 2025 Official leaderboard
68 Aya Expanse 32B 47.2% ±1.5 Oct 2024 Official leaderboard
69 Llama-3.1-Tulu-3-8B 46.5% ±1.1 Nov 2024 Official leaderboard
70 Gemma 3 4B 46.5% ±1.4 Mar 2025 Official leaderboard
71 GLM-4.5-Air 46.2% ±1.7 thinking Jul 2025 Official leaderboard
72 Qwen3.5-4B-Base 45.5% ±2.5 val Feb 2026 Official leaderboard
73 Reka Flash 3 44.9% ±1.2 Mar 2025 Official leaderboard
74 Voxtral Small 44.8% ±1.3 Jul 2025 Official leaderboard
75 Gemma 2 9B 44.7% ±2.1 Jun 2024 Official leaderboard
76 Apertus 8B Instruct 44.5% ±1.4 Sep 2025 Official leaderboard
77 Reka Flash 3.1 44.4% ±1.6 May 2025 Official leaderboard
78 Gemma 3 4B PT 44% ±1.8 Mar 2025 Official leaderboard
79 OLMo 2 32B Instruct 43.5% ±1.7 Mar 2025 Official leaderboard
80 deepseek-r1-0528-qwen3-8b 43.1% ±1.4 May 2025 Official leaderboard
81 Gemma 7B 42.4% ±1.9 Feb 2024 Official leaderboard
82 Ministral 3 14B 42.4% ±4.8 val Dec 2025 Official leaderboard
83 GPT-5.6 Luna 42.1% ±1.6 zero-shot, val Jul 2026 Official leaderboard
84 Qwen3.5-9B-Base 40.7% ±2.4 val Feb 2026 Official leaderboard
85 GLM-4.1V-9B-Thinking 40% ±1.2 Jul 2025 Official leaderboard
86 Qwen3-8B 40% ±1.8 Apr 2025 Official leaderboard
87 Qwen3-14B 38.5% ±4.6 Apr 2025 Official leaderboard
88 Gemma 4 E2B 37.9% ±3.8 val Mar 2026 Official leaderboard
89 DeepSeek-R1-Distill-Llama-70B 37.2% ±1.9 Jan 2025 Official leaderboard
90 gpt-oss-20b 37.2% ±1 medium Aug 2025 Official leaderboard
91 granite-4.0-micro 36.7% ±2.3 Oct 2025 Official leaderboard
92 nvidia-nemotron-nano-9b-v2 36.6% ±2.1 Aug 2025 Official leaderboard
93 Qwen3.5-2B 36.5% ±4.9 val Feb 2026 Official leaderboard
94 Llama 3.1-8B 36.3% ±2.3 Jul 2024 Official leaderboard
95 Qwen1.5-14B-Chat 35.6% ±1.1 Feb 2024 Official leaderboard
96 Qwen3-4B-Thinking-2507 35.4% ±1.9 Aug 2025 Official leaderboard
97 Ministral 3 3B Instruct 34.8% ±5.8 val Dec 2025 Official leaderboard
98 GPT-4.1 nano 34.7% ±1.9 zero-shot, val Apr 2025 Official leaderboard
99 Falcon3-10B-Instruct 34% ±2.6 Dec 2024 Official leaderboard
100 qwen3-4b-instruct-2507 32.5% ±1.5 Aug 2025 Official leaderboard
101 Olmo 3.1 32B Instruct 32.4% ±1.9 Dec 2025 Official leaderboard
102 Command R7B 32.2% ±1.6 Dec 2024 Official leaderboard
103 EXAONE 4.0 32B 31.6% ±3.9 val Jul 2025 Official leaderboard
104 Falcon3-7B-Instruct 31.5% ±1.6 Dec 2024 Official leaderboard
105 Granite-3.0-8B-Instruct 31.1% ±3.7 Oct 2024 Official leaderboard
106 Qwen3-4B 31% ±1.2 Apr 2025 Official leaderboard
107 Aya Expanse 8B 29.7% ±2.5 Oct 2024 Official leaderboard
108 Qwen3-1.7B 29.5% ±1.3 Apr 2025 Official leaderboard
109 Olmo 3.1 32B Think 29% ±4 val Dec 2025 Official leaderboard
110 AFM-4.5B 29% ±1.5 Jul 2025 Official leaderboard
111 Yi-1.5-34B 28.5% ±1.8 May 2024 Official leaderboard
112 Granite-3.0-8B-Base 27.6% ±2.4 Oct 2024 Official leaderboard
113 Ministral 3 3B Reasoning 27.4% ±1.4 Dec 2025 Official leaderboard
114 Mistral 7B v0.3 26.7% ±2.4 Sep 2023 Official leaderboard
115 Mistral 7B v0.1 25.5% ±3.1 Sep 2023 Official leaderboard
116 Gemma 2B 23.9% ±2.6 Feb 2024 Official leaderboard
117 Mistral-7B-Instruct-v0.2 23.8% ±1.8 Dec 2023 Official leaderboard
118 Gemma 3 1B PT 22.2% ±3.1 Mar 2025 Official leaderboard
119 DeepSeek-R1-Distill-Qwen-32B 21.3% ±3.5 Jan 2025 Official leaderboard
120 Qwen1.5-14B 21.1% ±0.8 Feb 2024 Official leaderboard
121 DeepSeek-R1-Distill-Qwen-14B 21.1% ±2 Jan 2025 Official leaderboard
122 Tiny Aya Global 21.1% ±5.4 val Feb 2026 Official leaderboard
123 Gemma 2 2B IT 20% ±1.8 Jul 2024 Official leaderboard
124 Qwen3.5-2B-Base 19.5% ±2.3 val Feb 2026 Official leaderboard
125 Tiny Aya Water 19.3% ±5.6 val Feb 2026 Official leaderboard
126 DeepSeek-R1-Distill-Qwen-7B 17.8% ±1.9 Jan 2025 Official leaderboard
127 Qwen3-0.6B 16.1% ±1.7 Apr 2025 Official leaderboard
128 Qwen3.5-0.8B 15% ±5.3 val Feb 2026 Official leaderboard
129 Gemma 3 1B 12.8% ±2 Mar 2025 Official leaderboard
130 LFM2-1.2B 12.5% ±2.1 Jul 2025 Official leaderboard
131 Mistral-7B-Instruct-v0.1 11.4% ±3.5 Sep 2023 Official leaderboard
132 Phi-4-reasoning 10.2% ±1.7 Apr 2025 Official leaderboard
133 LFM2.5-1.2B-Thinking 7.00% ±2.5 Jan 2026 Official leaderboard
134 LFM2.5-8B-A1B 6.08% ±7 val May 2026 Official leaderboard
135 SmolLM2 1.7B Instruct 5.78% ±1.9 Oct 2024 Official leaderboard
136 Qwen3.5-0.8B-Base 5.13% ±2.9 val Feb 2026 Official leaderboard
137 Gemma 3 270M 5.04% ±1.5 Aug 2025 Official leaderboard
138 Llama 3.2 1B 4.85% ±1.5 Sep 2024 Official leaderboard
139 Falcon3-3B-Instruct 3.87% ±2 Dec 2024 Official leaderboard
140 DeepSeek-R1-Distill-Qwen-1.5B 3.04% ±1.4 Jan 2025 Official leaderboard
141 SmolLM2-360M 2.97% ±1.7 Oct 2024 Official leaderboard
142 Phi-4-Reasoning-plus 2.32% ±1.6 Apr 2025 Official leaderboard
143 SmolLM2-135M 1.38% ±1.9 Oct 2024 Official leaderboard
144 SmolLM2-1.7B 0.77% ±2.1 Oct 2024 Official leaderboard
145 Phi-4-mini-reasoning 0.66% ±2.2 Apr 2025 Official leaderboard
146 Phi-2 -0.08% ±0.1 Dec 2023 Official leaderboard
147 LFM2.5-1.2B-Instruct -1.39% ±1.9 Jan 2026 Official leaderboard

Scores at each effort level

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
72.8%
Apertus 1.5 8B
zero-shot, … val
68.7%
Mistral Medium 3.5
zero-shot, … val
64.2%
Apertus 1.5 70B
zero-shot, … val
58.7%
gpt-oss-20b
low
37.2%
Ministral 3 3B Instruct
zero-shot, … val
34.8%
Aya Expanse 8B
val
29.7%

Source: Official FoSent leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Gemini 2.5 Pro Jun 2025 zero-shot, val 91.6%
Gemini 3 Pro zero-shot, val 83.1%
Gemini 3.7 Flash zero-shot, val 81.8%
o3 zero-shot, val 81%
GPT-6.2 Astra zero-shot, val 78%
Claude Sonnet 4.6 zero-shot, val 77.8%
GPT-4.1 zero-shot, val 77.8%
GPT-6 Astra zero-shot, val 77.7%
GPT-5 mini zero-shot, val 77.7%
GPT-5.6 Terra zero-shot, val 77.5%
Gemma 4 31B IT val 76.3%
GPT-6 Sol zero-shot, val 75.9%
Gemini 3.1 Flash-Lite zero-shot, val 75.3%
Gemini 3.6 Flash zero-shot, val 73.9%
Grok-2 Dec 2024 zero-shot, val 73.8%
GPT-6.1 Sol zero-shot, val 73.2%
GLM-5.3-Flash val 72.8%
Gemma 4 12B val 72.4%
Gemma 4 26B A4B val 71.4%
GPT-5.2 zero-shot, val 71.3%
GPT-6.2 Sol zero-shot, val 70.9%
Gemini 3.5 Flash-Lite zero-shot, val 70.3%
o4-mini zero-shot, val 69.6%
Apertus 1.5 8B val 68.7%
Gemini 2.5 Pro Jun 2025 zero-shot, val 68%
Claude 3.5 Haiku zero-shot, val 66%
glm-4.7-flash val 65.2%
GPT-5 zero-shot, val 65.1%
Qwen3.5-35B-A3B-Base val 64.9%
Mistral Medium 3.5 val 64.2%
Cogito v1 Preview 70B 63.9%
Claude Haiku 4.5 zero-shot, val 63.4%
Qwen3.5-9B val 63.3%
GPT-6 Luna zero-shot, val 62.5%
Gemini 1.5 Flash Sep 2024 zero-shot, val 62.1%
Qwen 3.8 27B val 62.1%
Grok 3 zero-shot, val 61.3%
Gemma 3 27B zero-shot 61.2%
Gemini 2.5 Flash-Lite Sep 2025 zero-shot, val 61.1%
GLM-5.3-Flash zero-shot, val 60.8%
Gemma 4 E4B val 60.2%
o3-mini zero-shot, val 59.9%
GPT-4.1 mini zero-shot, val 59.6%
GPT-5.6 Sol zero-shot, val 59.5%
Apertus 1.5 70B val 58.7%
Llama 3.1-405B 58.1%
Gemini 2.5 Flash Sep 2025 zero-shot, val 57%
Gemma 2 27B 56.5%
gpt-oss-120b val 56.4%
GPT-5 nano zero-shot, val 55.9%
Ministral 3 8B Instruct val 55.7%
Gemma 3 12B PT 55.4%
Apertus 1.5 8B zero-shot, val 54.9%
Qwen3-235B-A22B 54.4%
Gemma 3 12B 54.1%
Qwen3.5-4B val 53.9%
Llama 4 Scout 53.3%
Grok 4.1 Fast zero-shot, val 53.3%
Gemma 3n E4B 53%
Llama 3.1-70B 52.7%
Grok-3 mini zero-shot, val 52.6%
Granite 4.0 H Small 52.3%
Gemma 3 27B PT 52.2%
Mistral Medium 3.5 zero-shot, val 51.3%
Phi-4 51%
Qwen3-Next-80B-A3B-Instruct 50.4%
Nemotron-Cascade 2 30B-A3B val 50.2%
Gemma 3n E2B 50.2%
Apertus 70B 49.8%
Qwen3-Next-80B-A3B thinking 47.7%
QwQ-32B 47.7%
Gemma 3n E4B 47.3%
Aya Expanse 32B 47.2%
Llama-3.1-Tulu-3-8B 46.5%
Gemma 3 4B 46.5%
GLM-4.5-Air thinking 46.2%
Qwen3.5-4B-Base val 45.5%
Reka Flash 3 44.9%
Voxtral Small 44.8%
Gemma 2 9B 44.7%
Apertus 8B Instruct 44.5%
Reka Flash 3.1 44.4%
Gemma 3 4B PT 44%
OLMo 2 32B Instruct 43.5%
deepseek-r1-0528-qwen3-8b 43.1%
Gemma 7B 42.4%
Ministral 3 14B val 42.4%
Gemma 4 31B IT val 42.1%
GPT-5.6 Luna zero-shot, val 42.1%
Apertus 1.5 70B zero-shot, val 41%
Qwen3.5-9B-Base val 40.7%
GLM-4.1V-9B-Thinking 40%
Qwen3-8B 40%
Qwen3-14B 38.5%
Gemma 4 E2B val 37.9%
DeepSeek-R1-Distill-Llama-70B 37.2%
gpt-oss-20b medium 37.2%
granite-4.0-micro 36.7%
nvidia-nemotron-nano-9b-v2 36.6%
Qwen3.5-2B val 36.5%
Llama 3.1-8B 36.3%
Qwen1.5-14B-Chat 35.6%
Qwen3-4B-Thinking-2507 35.4%
Ministral 3 3B Instruct val 34.8%
GPT-4.1 nano zero-shot, val 34.7%
gpt-oss-20b high 34.6%
gpt-oss-20b low 34.5%
Falcon3-10B-Instruct 34%
Gemma 4 12B val 33.4%
qwen3-4b-instruct-2507 32.5%
Olmo 3.1 32B Instruct 32.4%
Command R7B 32.2%
EXAONE 4.0 32B val 31.6%
Falcon3-7B-Instruct 31.5%
Granite-3.0-8B-Instruct 31.1%
Qwen3-4B 31%
Gemma 4 26B A4B val 30.4%
Aya Expanse 8B 29.7%
Qwen3-1.7B 29.5%
Olmo 3.1 32B Think val 29%
AFM-4.5B 29%
Yi-1.5-34B 28.5%
Apertus 70B 28%
Granite-3.0-8B-Base 27.6%
Ministral 3 3B Reasoning 27.4%
Mistral 7B v0.3 26.7%
Mistral 7B v0.1 25.5%
Gemma 2B 23.9%
Mistral-7B-Instruct-v0.2 23.8%
Ministral 3 3B Instruct zero-shot, val 23%
Gemma 3 1B PT 22.2%
DeepSeek-R1-Distill-Qwen-32B 21.3%
Qwen1.5-14B 21.1%
DeepSeek-R1-Distill-Qwen-14B 21.1%
Tiny Aya Global val 21.1%
Gemma 2 2B IT 20%
Aya Expanse 8B val 20%
Qwen3.5-2B-Base val 19.5%
Tiny Aya Water val 19.3%
DeepSeek-R1-Distill-Qwen-7B 17.8%
Qwen3-0.6B 16.1%
Qwen3.5-0.8B val 15%
Gemma 4 E4B val 13.3%
Gemma 3 1B 12.8%
LFM2-1.2B 12.5%
Mistral-7B-Instruct-v0.1 11.4%
Phi-4-reasoning 10.2%
Gemma 3 1B 9.18%
LFM2.5-1.2B-Thinking 7.00%
LFM2.5-8B-A1B val 6.08%
SmolLM2 1.7B Instruct 5.78%
Qwen3.5-0.8B-Base val 5.13%
Gemma 3 270M 5.04%
Llama 3.2 1B 4.85%
Falcon3-3B-Instruct 3.87%
DeepSeek-R1-Distill-Qwen-1.5B 3.04%
SmolLM2-360M 2.97%
Phi-4-Reasoning-plus 2.32%
SmolLM2-135M 1.38%
SmolLM2-1.7B 0.77%
Phi-4-mini-reasoning 0.66%
Gemma 3 270M 0.03%
Phi-2 -0.08%
LFM2.5-1.2B-Instruct -1.39%
Gemma 4 E2B val -1.68%

Scores at each effort level

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
72.8%
Apertus 1.5 8B
zero-shot, … val
68.7%
Mistral Medium 3.5
zero-shot, … val
64.2%
Apertus 1.5 70B
zero-shot, … val
58.7%
gpt-oss-20b
low
37.2%
Ministral 3 3B Instruct
zero-shot, … val
34.8%
Aya Expanse 8B
val
29.7%

Source: Official FoSent leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Gemini 2.5 Pro Jun 2025 zero-shot, val 91.6%
Gemini 3 Pro zero-shot, val 83.1%
Gemini 3.7 Flash zero-shot, val 81.8%
o3 zero-shot, val 81%
GPT-6.2 Astra zero-shot, val 78%
Claude Sonnet 4.6 zero-shot, val 77.8%
GPT-4.1 zero-shot, val 77.8%
GPT-6 Astra zero-shot, val 77.7%
GPT-5 mini zero-shot, val 77.7%
GPT-5.6 Terra zero-shot, val 77.5%
Gemma 4 31B IT val 76.3%
GPT-6 Sol zero-shot, val 75.9%
Gemini 3.1 Flash-Lite zero-shot, val 75.3%
Gemini 3.6 Flash zero-shot, val 73.9%
Grok-2 Dec 2024 zero-shot, val 73.8%
GPT-6.1 Sol zero-shot, val 73.2%
GLM-5.3-Flash val 72.8%
Gemma 4 12B val 72.4%
Gemma 4 26B A4B val 71.4%
GPT-5.2 zero-shot, val 71.3%
GPT-6.2 Sol zero-shot, val 70.9%
Gemini 3.5 Flash-Lite zero-shot, val 70.3%
o4-mini zero-shot, val 69.6%
Apertus 1.5 8B val 68.7%
Gemini 2.5 Pro Jun 2025 zero-shot, val 68%
Claude 3.5 Haiku zero-shot, val 66%
glm-4.7-flash val 65.2%
GPT-5 zero-shot, val 65.1%
Qwen3.5-35B-A3B-Base val 64.9%
Mistral Medium 3.5 val 64.2%
Cogito v1 Preview 70B 63.9%
Claude Haiku 4.5 zero-shot, val 63.4%
Qwen3.5-9B val 63.3%
GPT-6 Luna zero-shot, val 62.5%
Gemini 1.5 Flash Sep 2024 zero-shot, val 62.1%
Qwen 3.8 27B val 62.1%
Grok 3 zero-shot, val 61.3%
Gemma 3 27B zero-shot 61.2%
Gemini 2.5 Flash-Lite Sep 2025 zero-shot, val 61.1%
GLM-5.3-Flash zero-shot, val 60.8%
Gemma 4 E4B val 60.2%
o3-mini zero-shot, val 59.9%
GPT-4.1 mini zero-shot, val 59.6%
GPT-5.6 Sol zero-shot, val 59.5%
Apertus 1.5 70B val 58.7%
Llama 3.1-405B 58.1%
Gemini 2.5 Flash Sep 2025 zero-shot, val 57%
Gemma 2 27B 56.5%
gpt-oss-120b val 56.4%
GPT-5 nano zero-shot, val 55.9%
Ministral 3 8B Instruct val 55.7%
Gemma 3 12B PT 55.4%
Apertus 1.5 8B zero-shot, val 54.9%
Qwen3-235B-A22B 54.4%
Gemma 3 12B 54.1%
Qwen3.5-4B val 53.9%
Llama 4 Scout 53.3%
Grok 4.1 Fast zero-shot, val 53.3%
Gemma 3n E4B 53%
Llama 3.1-70B 52.7%
Grok-3 mini zero-shot, val 52.6%
Granite 4.0 H Small 52.3%
Gemma 3 27B PT 52.2%
Mistral Medium 3.5 zero-shot, val 51.3%
Phi-4 51%
Qwen3-Next-80B-A3B-Instruct 50.4%
Nemotron-Cascade 2 30B-A3B val 50.2%
Gemma 3n E2B 50.2%
Apertus 70B 49.8%
Qwen3-Next-80B-A3B thinking 47.7%
QwQ-32B 47.7%
Gemma 3n E4B 47.3%
Aya Expanse 32B 47.2%
Llama-3.1-Tulu-3-8B 46.5%
Gemma 3 4B 46.5%
GLM-4.5-Air thinking 46.2%
Qwen3.5-4B-Base val 45.5%
Reka Flash 3 44.9%
Voxtral Small 44.8%
Gemma 2 9B 44.7%
Apertus 8B Instruct 44.5%
Reka Flash 3.1 44.4%
Gemma 3 4B PT 44%
OLMo 2 32B Instruct 43.5%
deepseek-r1-0528-qwen3-8b 43.1%
Gemma 7B 42.4%
Ministral 3 14B val 42.4%
Gemma 4 31B IT val 42.1%
GPT-5.6 Luna zero-shot, val 42.1%
Apertus 1.5 70B zero-shot, val 41%
Qwen3.5-9B-Base val 40.7%
GLM-4.1V-9B-Thinking 40%
Qwen3-8B 40%
Qwen3-14B 38.5%
Gemma 4 E2B val 37.9%
DeepSeek-R1-Distill-Llama-70B 37.2%
gpt-oss-20b medium 37.2%
granite-4.0-micro 36.7%
nvidia-nemotron-nano-9b-v2 36.6%
Qwen3.5-2B val 36.5%
Llama 3.1-8B 36.3%
Qwen1.5-14B-Chat 35.6%
Qwen3-4B-Thinking-2507 35.4%
Ministral 3 3B Instruct val 34.8%
GPT-4.1 nano zero-shot, val 34.7%
gpt-oss-20b high 34.6%
gpt-oss-20b low 34.5%
Falcon3-10B-Instruct 34%
Gemma 4 12B val 33.4%
qwen3-4b-instruct-2507 32.5%
Olmo 3.1 32B Instruct 32.4%
Command R7B 32.2%
EXAONE 4.0 32B val 31.6%
Falcon3-7B-Instruct 31.5%
Granite-3.0-8B-Instruct 31.1%
Qwen3-4B 31%
Gemma 4 26B A4B val 30.4%
Aya Expanse 8B 29.7%
Qwen3-1.7B 29.5%
Olmo 3.1 32B Think val 29%
AFM-4.5B 29%
Yi-1.5-34B 28.5%
Apertus 70B 28%
Granite-3.0-8B-Base 27.6%
Ministral 3 3B Reasoning 27.4%
Mistral 7B v0.3 26.7%
Mistral 7B v0.1 25.5%
Gemma 2B 23.9%
Mistral-7B-Instruct-v0.2 23.8%
Ministral 3 3B Instruct zero-shot, val 23%
Gemma 3 1B PT 22.2%
DeepSeek-R1-Distill-Qwen-32B 21.3%
Qwen1.5-14B 21.1%
DeepSeek-R1-Distill-Qwen-14B 21.1%
Tiny Aya Global val 21.1%
Gemma 2 2B IT 20%
Aya Expanse 8B val 20%
Qwen3.5-2B-Base val 19.5%
Tiny Aya Water val 19.3%
DeepSeek-R1-Distill-Qwen-7B 17.8%
Qwen3-0.6B 16.1%
Qwen3.5-0.8B val 15%
Gemma 4 E4B val 13.3%
Gemma 3 1B 12.8%
LFM2-1.2B 12.5%
Mistral-7B-Instruct-v0.1 11.4%
Phi-4-reasoning 10.2%
Gemma 3 1B 9.18%
LFM2.5-1.2B-Thinking 7.00%
LFM2.5-8B-A1B val 6.08%
SmolLM2 1.7B Instruct 5.78%
Qwen3.5-0.8B-Base val 5.13%
Gemma 3 270M 5.04%
Llama 3.2 1B 4.85%
Falcon3-3B-Instruct 3.87%
DeepSeek-R1-Distill-Qwen-1.5B 3.04%
SmolLM2-360M 2.97%
Phi-4-Reasoning-plus 2.32%
SmolLM2-135M 1.38%
SmolLM2-1.7B 0.77%
Phi-4-mini-reasoning 0.66%
Gemma 3 270M 0.03%
Phi-2 -0.08%
LFM2.5-1.2B-Instruct -1.39%
Gemma 4 E2B val -1.68%

9 models have held number one on the official leaderboard since its release in May 2024

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

Google DeepMind Meta Anthropic xAI OpenAI
Own leaderboard Passed 85% in Jun 2025, by Gemini 2.5 Pro

Source: Official FoSent leaderboard

Gemini 2.5 Pro on the official leaderboard passed 85% in June 2025

That was 13 months after its release. By the same age, 24% of the benchmarks from 2024 had passed 85%. 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%) Gemini 2.5 Pro o3 Grok-2 Claude 3.5 Haiku Llama 3.1-405B Gemma 7B Qwen1.5-14B-Chat 0% 20% 40% 60% 80% 100% 2024 2025 2026 2027 Near the ceiling (85%) Gemini 2.5 Pro o3

Source: Official FoSent 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 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (540 entries).
  • Pass Models scored 147 models we track, plus 393 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).
  • Note Headroom Saturated. The best is 91.6%, past 85% since June 2025.
Leaderboard import 3 pass
  • Pass Importer We read the board on euroeval.com: 165 rows for models we track and 393 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