Benchmark Almanac

Dutch Book Reviews Dataset

Qwen/Qwen3-235B-A22B-Instruct-2507-FP8 93% tops its leaderboard OLMo 2 32B Instruct 92.5% leads the models we track
Not in model cards

Models classify the sentiment of Dutch book reviews from Hebban using ratings supplied by readers.

EuroEval scores sentiment labels with the Matthews correlation coefficient.

Released
October 2019
Kind
Reading in many languages
Level
Everyday
Scored
Checked automatically
Answers
Multiple choice
Tasks
All public
Measures
Capability
Language
Dutch
In and out
Text → Text
Interaction
Single turn
Items from
Real work and use
Refresh
One fixed set
Run by
Anyone
Strongest model tested
Qwen3-235B-A22B #131 on the capability index
Last model added
March 2026
Scores refreshed
November 20th, 2026
352
Entries on EuroEval’s leaderboard, 105 models we track

Qwen/Qwen3-235B-A22B-Instruct-2507-FP8 tops the leaderboard with 93% OLMo 2 32B Instruct leads with 92.5% among models released by March 2026

Best score per entry from EuroEval’s leaderboard, read on November 20th, 2026. EuroEval reports Matthews correlation on its DBRD-mini test split, scaled by 100. 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. The best model we track is OLMo 2 32B Instruct, at 92.5%. This view shows only models we track, each at its best.

The newest model with a score here came out in March 2026, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:

352 entries. Scroll sideways for the rest.

  1. 93.0 Qwen/Qwen3-235B-A22B-Instruct-2507-FP8
  2. 92.5 OLMo 2 32B Instruct
  3. 92.5 Gemma 3 27B
  4. 92.3 Llama 3.1-405B
  5. 92.3 NovaSky-AI/Sky-T1-32B-Preview
  6. 92.1 Qwen3-Next-80B-A3B-Instruct
  7. 92.1 Llama 3.3 70B Instruct
  8. 92.0 ThatsGroes/gemma-2-27b-it-FP8-Dynamic
  9. 91.9 Apertus 70B
  10. 91.9 Mistral Small 3.1
  11. 91.9 Llama 3-8B
  12. 91.8 Qwen3-14B
  13. 91.7 Aya Expanse 8B
  14. 91.7 Qwen3-32B
  15. 91.7 Llama 3.1-8B
  16. 91.6 allenai/Olmo-3.1-32B-Instruct-DPO
  17. 91.5 Llama-3.1-Tulu-3-8B
  18. 91.5 allenai/Llama-3.1-Tulu-3-8B-DPO
  19. 91.4 Olmo 3.1 32B Instruct
  20. 91.4 Qwen/Qwen3-235B-A22B-Thinking-2507-FP8
  21. 91.4 Mistral Small 3.2
  22. 91.4 01-ai/Yi-1.5-9B
  23. 91.4 ai21labs/Jamba-v0.1
  24. 91.4 allenai/Llama-3.1-Tulu-3.1-8B
  25. 91.3 Aya Expanse 32B
  26. 91.3 Cogito v1 Preview 70B
  27. 91.3 google/gemma-2-9b-it
  28. 91.3 Llama-3.1-Nemotron-70B-Instruct
  29. 91.3 google/gemma-2-27b-it
  30. 91.2 LumiOpen/Llama-Poro-2-70B-SFT
  31. 91.2 Gemma 3 12B
  32. 91.1 mistralai/Magistral-Small-2509
  33. 91.1 deepcogito/cogito-v2-preview-llama-70B
  34. 91.1 meta-llama/Llama-3.1-405B-Instruct-FP8
  35. 91.0 Mixtral 8x7B
  36. 91.0 deepcogito/cogito-v1-preview-qwen-32B
  37. 90.9 meta-llama/Llama-4-Scout-17B-16E
  38. 90.8 GLM-4.5-Air
  39. 90.8 QwQ-32B
  40. 90.8 meta-llama/Llama-3.1-70B-Instruct
  41. 90.7 Qwen3-30B-A3B
  42. 90.7 mistralai/Mistral-Small-24B-Instruct-2501
  43. 90.7 Reka Flash 3
  44. 90.7 Qwen2.5-72B
  45. 90.7 Voxtral Small
  46. 90.7 Llama 3-70B
  47. 90.7 Gemma 3 12B PT
  48. 90.6 utter-project/EuroLLM-22B-Instruct-Preview
  49. 90.5 speakleash/Bielik-11B-v2.3-Instruct
  50. 90.5 Qwen3-Next-80B-A3B
  51. 90.5 Mistral 7B v0.3
  52. 90.5 tiiuae/falcon-11B
  53. 90.5 Llama 2-70B
  54. 90.4 allenai/Olmo-3.1-32B-Instruct-SFT
  55. 90.4 utter-project/EuroLLM-22B-Instruct-2512
  56. 90.3 bullerwins/Magistral-Small-2506-fp8
  57. 90.3 utter-project/EuroLLM-22B-Preview
  58. 90.2 Gemma 2 9B
  59. 90.2 deepcogito/cogito-v1-preview-llama-8B
  60. 90.2 meta-llama/Llama-2-13b-hf
  61. 90.2 Command R (03-2024)
  62. 90.2 Granite 4.0 H Small
  63. 90.2 Llama 3.1-70B
  64. 90.2 GLM-4.1V-9B-Thinking
  65. 90.1 Gemma 3 27B PT
  66. 90.1 meta-llama/Meta-Llama-3-70B-Instruct
  67. 90.1 c4ai-command-r-08-2024
  68. 90.0 Qwen3-235B-A22B
  69. 90.0 Nemotron 3 Nano 30B-A3B
  70. 90.0 ChocoLlama/Llama-3-ChocoLlama-8B-instruct
  71. 90.0 meta-llama/Llama-3.1-8B-Instruct
  72. 90.0 deepcogito/cogito-v1-preview-qwen-14B
  73. 89.9 occiglot/occiglot-7b-eu5
  74. 89.9 BSC-LT/salamandra-7b
  75. 89.9 Reka Flash 3.1
  76. 89.9 Yi-1.5-34B
  77. 89.8 LumiOpen/Llama-Poro-2-70B-Instruct
  78. 89.8 DeepSeek-R1-Distill-Llama-70B
  79. 89.7 LumiOpen/Viking-33B
  80. 89.7 HuggingFaceTB/SmolLM3-3B-Base
  81. 89.7 Phi-4
  82. 89.7 Llama 4 Scout
  83. 89.6 Qwen3-8B
  84. 89.5 Qwen1.5-14B-Chat
  85. 89.3 allenai/OLMo-2-0325-32B
  86. 89.3 ibm-granite/granite-4.0-h-tiny
  87. 89.3 occiglot/occiglot-7b-eu5-instruct
  88. 89.2 mistralai/Mixtral-8x7B-Instruct-v0.1
  89. 89.2 152334H/miqu-1-70b-sf
  90. 89.2 ibm-granite/granite-3.0-3b-a800m-base
  91. 89.2 meta-llama/Meta-Llama-3-8B-Instruct
  92. 89.1 ibm-granite/granite-4.0-h-micro
  93. 89.1 Granite-3.0-2B-Base
  94. 89.0 utter-project/EuroLLM-9B-Instruct
  95. 89.0 Gemma 3n E2B
  96. 89.0 CohereForAI/aya-23-8B
  97. 89.0 mistralai/Ministral-3-8B-Reasoning-2512
  98. 89.0 mistralai/Mistral-Nemo-Instruct-2407
  99. 89.0 mistralai/Ministral-8B-Instruct-2410
  100. 88.9 meta-llama/Llama-2-7b-hf
  101. 88.9 typesafe/jev-1.13.0
  102. 88.8 allenai/Olmo-3-1025-7B
  103. 88.8 allenai/Olmo-3-1125-32B
  104. 88.8 Llama 3.2 3B
  105. 88.8 DeepSeek-R1-Distill-Qwen-32B
  106. 88.7 allenai/OLMo-2-1124-13B-Instruct
  107. 88.6 Gemma 3n E4B
  108. 88.5 DeepSeek-R1-Distill-Qwen-14B
  109. 88.5 Ministral 3 3B Reasoning
  110. 88.5 utter-project/EuroLLM-22B-2512
  111. 88.5 BramVanroy/GEITje-7B-ultra
  112. 88.4 ibm-granite/granite-3.0-3b-a800m-instruct
  113. 88.4 Qwen3-4B-Thinking-2507
  114. 88.4 NousResearch/DeepHermes-3-Llama-3-8B-Preview
  115. 88.4 deepseek-r1-0528-qwen3-8b
  116. 88.3 Gemma 7B
  117. 88.2 allenai/Llama-3.1-Tulu-3-8B-SFT
  118. 88.1 Granite-3.0-8B-Instruct
  119. 88.1 mistralai/Mistral-Small-3.1-24B-Base-2503
  120. 88.0 Mistral 7B v0.1
  121. 88.0 allenai/OLMo-2-1124-13B
  122. 88.0 neuralmagic/Sparse-Llama-3.1-8B-2of4
  123. 88.0 Granite-3.0-8B-Base
  124. 87.9 gpt-oss-20b
  125. 87.8 norallm/normistral-11b-warm
  126. 87.8 upstage/SOLAR-10.7B-v1.0
  127. 87.8 01-ai/Yi-1.5-6B
  128. 87.7 Falcon3-7B-Instruct
  129. 87.6 CohereForAI/aya-23-35B
  130. 87.6 nvidia/mistral-nemo-minitron-8b-base
  131. 87.5 swiss-ai/Apertus-8B-2509
  132. 87.5 ibm-granite/granite-34b-code-base
  133. 87.5 Qwen1.5-14B
  134. 87.5 google/gemma-2-2b
  135. 87.4 mistralai/Voxtral-Mini-3B-2507
  136. 87.2 granite-4.0-micro
  137. 87.1 meta-llama/Llama-3.2-3B-Instruct
  138. 86.9 utter-project/EuroLLM-9B
  139. 86.9 tiiuae/Falcon3-7B-Base
  140. 86.7 deepcogito/cogito-v1-preview-llama-3B
  141. 86.6 nvidia/mistral-nemo-minitron-8b-instruct
  142. 86.6 Gemma 3 4B
  143. 86.6 HuggingFaceTB/SmolLM3-3B
  144. 86.5 mistralai/Ministral-3-3B-Base-2512
  145. 86.5 Gemma 2B
  146. 86.5 Mistral-7B-Instruct-v0.2
  147. 86.3 qwen3-4b-instruct-2507
  148. 86.2 TrustLLMeu/baseline-7-8b_2-3t-tokens_llama
  149. 86.1 Granite-3.0-2B-Instruct
  150. 86.1 Gemma 3 4B PT
  151. 86.1 BSC-LT/salamandra-7b-instruct
  152. 86.0 mii-llm/nesso-4B
  153. 85.9 allenai/OLMo-2-1124-7B-Instruct
  154. 85.8 stabilityai/stablelm-2-1_6b
  155. 85.8 speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551
  156. 85.6 Starling-LM-7B-beta
  157. 85.6 nvidia-nemotron-nano-9b-v2
  158. 85.4 ibm-granite/granite-34b-code-instruct
  159. 85.3 speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c
  160. 85.3 Command R7B
  161. 85.3 Gemma 2 27B
  162. 85.3 intfloat/multilingual-e5-large
  163. 85.3 ibm-granite/granite-20b-code-base
  164. 85.2 NorwAI/NorwAI-Mistral-7B
  165. 85.1 allenai/OLMo-2-1124-7B
  166. 85.1 TildeAI/TildeOpen-30b
  167. 85.0 ollama_chat/gemma3:4b-it-q4_K_M
  168. 85.0 google/rembert
  169. 84.9 Gemma 2 2B IT
  170. 84.8 AFM-4.5B
  171. 84.8 baidu/ERNIE-4.5-21B-A3B-PT
  172. 84.7 senseable/WestLake-7B-v2
  173. 84.6 LumiOpen/Viking-13B
  174. 84.5 speakleash/Bielik-11B-v2.6-Instruct
  175. 84.4 meta-llama/Llama-2-13b-chat-hf
  176. 84.3 Qwen3-4B
  177. 84.2 meta-llama/Llama-2-70b-chat-hf
  178. 84.2 Apertus 8B Instruct
  179. 84.1 TrustLLMeu/baseline-7-8b_1t-tokens_llama
  180. 83.9 MaLA-LM/emma-500-llama2-7b
  181. 83.8 TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49
  182. 83.7 Qwen/Qwen1.5-4B
  183. 83.7 Mistral-7B-Instruct-v0.1
  184. 83.2 ollama_chat/hf.co/google/gemma-3-4b-it-qat-q4_0-gguf
  185. 82.9 Trinity Mini
  186. 82.9 arcee-ai/AFM-4.5B-Base
  187. 82.9 Qwen3-1.7B
  188. 82.8 ChocoLlama/ChocoLlama-2-7B-base
  189. 82.7 Qwen/Qwen2.5-Math-7B
  190. 82.4 ibm-granite/granite-7b-base
  191. 82.3 Qwen1.5-4B-Chat
  192. 81.5 ChocoLlama/Llama-3-ChocoLlama-8B-base
  193. 81.0 EuroBERT/EuroBERT-610m
  194. 80.6 Llama 3.2 1B
  195. 80.1 Qwen/Qwen2.5-1.5B-Instruct
  196. 80.1 ibm-granite/granite-8b-code-base-4k
  197. 80.0 jhu-clsp/mmBERT-base
  198. 79.9 allenai/Olmo-3-7B-Instruct
  199. 79.8 allenai/Olmo-3-7B-Instruct-DPO
  200. 79.6 allenai/Olmo-3-7B-Think
  201. 79.5 ibm-granite/granite-8b-code-instruct-4k
  202. 79.4 Phi-3-mini-128k-instruct
  203. 79.1 allenai/Olmo-3-7B-Instruct-SFT
  204. 78.7 Nemotron 3 Nano 4B
  205. 78.7 LiquidAI/LFM2-2.6B
  206. 78.1 OmniJev/OneJev1.6-0.8B
  207. 77.2 setu4993/LaBSE
  208. 77.0 deepseek-ai/DeepSeek-R1-Distill-Llama-8B
  209. 77.0 intfloat/multilingual-e5-base
  210. 75.2 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6
  211. 74.9 Hunyuan-4B-Instruct
  212. 74.9 ZurichNLP/unsup-simcse-xlm-roberta-base
  213. 74.8 syvai/claes-preview
  214. 74.6 EuroBERT/EuroBERT-210m
  215. 74.5 timpal0l/Mistral-7B-v0.1-flashback-v2
  216. 74.4 DeepSeek-R1-Distill-Qwen-7B
  217. 74.3 SmolLM2-1.7B
  218. 74.1 SmolLM2 1.7B Instruct
  219. 74.1 Gemma 3 1B
  220. 73.4 ibm-granite/granite-3.0-1b-a400m-base
  221. 73.2 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f
  222. 72.2 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660
  223. 71.4 TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894
  224. 71.1 AI-Sweden-Models/ModernBERT-large
  225. 71.0 common-pile/comma-v0.1-2t
  226. 70.8 convaiinnovations/laya-multilingual
  227. 70.7 ibm-granite/granite-20b-code-instruct
  228. 70.2 HPLT/gpt-13b-nordic-prerelease
  229. 70.1 jaredpalmer/kev-0.5b
  230. 70.0 state-spaces/mamba-2.8b-hf
  231. 69.4 infly/OpenCoder-8B-Instruct
  232. 69.2 jhu-clsp/mmBERT-small
  233. 68.8 HPLT/hplt2c_nld_checkpoints
  234. 68.5 ChocoLlama/ChocoLlama-2-7B-tokentrans-base
  235. 68.3 FacebookAI/xlm-roberta-large
  236. 68.0 speakleash/Bielik-11B-v3-Base-20250730
  237. 67.7 infly/OpenCoder-8B-Base
  238. 67.1 Qwen/Qwen1.5-1.8B-Chat
  239. 66.9 ChocoLlama/ChocoLlama-2-7B-instruct
  240. 66.5 utter-project/EuroMoE-2.6B-A0.6B-Preview
  241. 66.4 tencent/Hunyuan-7B-Instruct
  242. 65.8 arcee-ai/Trinity-Mini-Base
  243. 65.8 syvai/claes-v1
  244. 65.5 LFM2.5-1.2B-Instruct
  245. 65.1 common-pile/comma-v0.1-1t
  246. 65.0 fastino/GLiNER2.5-Decide
  247. 64.6 ibm-granite/granite-3.0-1b-a400m-instruct
  248. 64.1 Qwen3-0.6B
  249. 63.1 openGPT-X/Teuken-7B-instruct-research-v0.4
  250. 62.2 yulan-team/YuLan-Mini-Instruct
  251. 61.5 microsoft/mdeberta-v3-base
  252. 61.0 Falcon3-3B-Instruct
  253. 60.1 openGPT-X/Teuken-7B-instruct-commercial-v0.4
  254. 59.6 sentence-transformers/stsb-xlm-r-multilingual
  255. 58.9 google/gemma-2b-it
  256. 58.1 LiquidAI/LFM2.5-1.2B-Base
  257. 58.0 facebook/MobileLLM-R1-950M-base
  258. 56.4 Gemma 3 1B PT
  259. 55.1 Twitter/twhin-bert-base
  260. 54.9 arcee-ai/Trinity-Nano-Base
  261. 53.2 swiss-ai/Apertus-v1.1-1.5B
  262. 53.2 PleIAs/Pleias-3b-Preview
  263. 53.1 BSC-LT/ALIA-40b
  264. 52.4 meta-llama/Llama-2-7b-chat-hf
  265. 51.7 Phi-4-mini-reasoning
  266. 49.6 ibm-granite/granite-3b-code-base-2k
  267. 49.3 google/gemma-7b-it
  268. 49.2 RuterNorway/Llama-2-7b-chat-norwegian
  269. 48.3 sentence-transformers/paraphrase-xlm-r-multilingual-v1
  270. 48.2 PleIAs/Pleias-Nano
  271. 47.7 microsoft/xlm-align-base
  272. 46.9 tiiuae/Falcon3-3B-Base
  273. 45.0 NorwAI/NorwAI-Mixtral-8x7B-instruct
  274. 44.8 speakleash/Bielik-4.5B-v3.0-Instruct
  275. 44.6 tencent/Hunyuan-1.8B-Instruct
  276. 43.0 Phi-4-reasoning
  277. 40.3 sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2
  278. 38.7 Falcon3-1B-Instruct
  279. 38.5 sentence-transformers/distiluse-base-multilingual-cased-v1
  280. 37.9 LFM2-1.2B
  281. 37.6 PleIAs/Pleias-1.2b-Preview
  282. 37.2 sentence-transformers/distiluse-base-multilingual-cased
  283. 37.2 sentence-transformers/distiluse-base-multilingual-cased-v2
  284. 37.2 speakleash/Bielik-1.5B-v3.0-Instruct
  285. 35.0 distilbert/distilbert-base-multilingual-cased
  286. 33.9 DeepSeek-R1-Distill-Qwen-1.5B
  287. 33.9 sentence-transformers/quora-distilbert-multilingual
  288. 33.8 sentence-transformers/distilbert-multilingual-nli-stsb-quora-ranking
  289. 33.5 answerdotai/ModernBERT-base
  290. 33.1 LFM2.5-1.2B-Thinking
  291. 32.8 AI-Sweden-Models/gpt-sw3-20b
  292. 31.2 Twitter/twhin-bert-large
  293. 30.1 AI-Sweden-Models/gpt-sw3-20b-instruct
  294. 29.5 EuropeanParliament/EUBERT
  295. 29.5 Geotrend/distilbert-base-25lang-cased
  296. 28.3 answerdotai/ModernBERT-large
  297. 27.5 AI-Sweden-Models/roberta-large-1160k
  298. 27.5 facebook/MobileLLM-R1-360M-base
  299. 27.3 Qwen/Qwen1.5-1.8B
  300. 27.0 arcee-ai/Trinity-Nano-Preview
  301. 26.7 Qwen/Qwen2.5-Math-1.5B
  302. 25.5 infly/OpenCoder-1.5B-Instruct
  303. 25.4 cardiffnlp/twitter-xlm-roberta-base
  304. 24.2 Phi-2
  305. 23.9 AI-Sweden-Models/bert-large-nordic-pile-1M-steps
  306. 23.6 ibm-granite/granite-3b-code-instruct-2k
  307. 22.1 convaiinnovations/laya
  308. 21.8 Gemma 3 270M
  309. 17.8 meta-llama/Llama-3.2-1B-Instruct
  310. 17.5 Phi-4-Reasoning-plus
  311. 14.9 HPLT/hplt2c_eng_checkpoints
  312. 14.6 SmolLM2-360M
  313. 13.7 LiquidAI/LFM2-350M
  314. 13.6 dbmdz/bert-base-historic-multilingual-cased
  315. 13.2 LiquidAI/LFM2-700M
  316. 13.2 Qwen/Qwen1.5-0.5B-Chat
  317. 10.9 Qwen/Qwen1.5-0.5B
  318. 10.3 utter-project/EuroVLM-1.7B-Preview
  319. 9.67 PleIAs/Pleias-Pico
  320. 9.54 fresh-xlm-roberta-base
  321. 8.96 PleIAs/Pleias-350m-Preview
  322. 7.64 baidu/ERNIE-4.5-0.3B-PT
  323. 6.87 HPLT/hplt2c_deu_checkpoints
  324. 6.81 facebook/MobileLLM-R1-140M-base
  325. 5.61 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview
  326. 4.28 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base
  327. 3.50 DiscoResearch/DiscoLM-70b
  328. 2.62 AI-Sweden-Models/gpt-sw3-356m-instruct
  329. 2.13 dbmdz/bert-tiny-historic-multilingual-cased
  330. 1.78 SmolLM2-135M
  331. 0.91 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct
  332. 0.86 PleIAs/Baguettotron@refs/pr/6
  333. 0.67 AI-Sweden-Models/gpt-sw3-126m-instruct
  334. 0.38 AI-Sweden-Models/gpt-sw3-356m
  335. 0.23 fresh-electra-small
  336. 0.23 infly/OpenCoder-1.5B-Base
  337. 0.06 facebook/MobileLLM-R1-360M
  338. 0.0 ChocoLlama/ChocoLlama-2-7B-tokentrans-instruct
  339. 0.0 facebook/MobileLLM-R1-950M
  340. 0.0 kyutai/helium-1-2b
  341. 0.0 openai-community/gpt2
  342. 0.0 openai-community/gpt2-large
  343. 0.0 openai-community/gpt2-medium
  344. 0.0 Shekswess/trlm-135m
  345. -0.76 facebook/MobileLLM-R1-140M
  346. -0.98 HuggingFaceTB/SmolLM2-135M-Instruct
  347. -1.28 PleIAs/Monad
  348. -1.60 AI-Sweden-Models/gpt-sw3-126m
  349. -2.34 ibm-granite/granite-7b-instruct
  350. -4.94 ssmits/Falcon2-5.5B-multilingual
  351. -6.52 HuggingFaceTB/SmolLM2-360M-Instruct
  352. -13.48 RJuro/kanelsnegl-v0.1
  1. 1 Qwen/Qwen3-235B-A22B-Instruct-2507-FP8 93%
  2. 2 OLMo 2 32B Instruct 92.5%
  3. 3 Gemma 3 27B 92.5%
  4. 4 Llama 3.1-405B 92.3%
  5. 5 NovaSky-AI/Sky-T1-32B-Preview 92.3%
  6. 6 Qwen3-Next-80B-A3B-Instruct 92.1%
  7. 7 Llama 3.3 70B Instruct 92.1%
  8. 8 ThatsGroes/gemma-2-27b-it-FP8-Dynamic 92%
  9. 9 Apertus 70B 91.9%
  10. 10 Mistral Small 3.1 91.9%
  11. 11 Llama 3-8B 91.9%
  12. 12 Qwen3-14B no-thinking 91.8%
  13. 13 Aya Expanse 8B 91.7%
  14. 14 Qwen3-32B no-thinking 91.7%
  15. 15 Llama 3.1-8B 91.7%
  16. 16 allenai/Olmo-3.1-32B-Instruct-DPO 91.6%
  17. 17 Llama-3.1-Tulu-3-8B 91.5%
  18. 18 allenai/Llama-3.1-Tulu-3-8B-DPO 91.5%
  19. 19 Olmo 3.1 32B Instruct 91.4%
  20. 20 Qwen/Qwen3-235B-A22B-Thinking-2507-FP8 91.4%
  21. 21 Mistral Small 3.2 91.4%
  22. 22 01-ai/Yi-1.5-9B 91.4%
  23. 23 ai21labs/Jamba-v0.1 91.4%
  24. 24 allenai/Llama-3.1-Tulu-3.1-8B 91.4%
  25. 25 Aya Expanse 32B 91.3%
  26. 26 Cogito v1 Preview 70B 91.3%
  27. 27 google/gemma-2-9b-it 91.3%
  28. 28 Llama-3.1-Nemotron-70B-Instruct 91.3%
  29. 29 google/gemma-2-27b-it 91.3%
  30. 30 LumiOpen/Llama-Poro-2-70B-SFT 91.2%
  31. 31 Gemma 3 12B 91.2%
  32. 32 mistralai/Magistral-Small-2509 91.1%
  33. 33 deepcogito/cogito-v2-preview-llama-70B 91.1%
  34. 34 meta-llama/Llama-3.1-405B-Instruct-FP8 91.1%
  35. 35 Mixtral 8x7B 91%
  36. 36 deepcogito/cogito-v1-preview-qwen-32B 91%
  37. 37 meta-llama/Llama-4-Scout-17B-16E 90.9%
  38. 38 GLM-4.5-Air thinking 90.8%
  39. 39 QwQ-32B 90.8%
  40. 40 meta-llama/Llama-3.1-70B-Instruct 90.8%
  41. 41 Qwen3-30B-A3B no-thinking 90.7%
  42. 42 mistralai/Mistral-Small-24B-Instruct-2501 90.7%
  43. 43 Reka Flash 3 90.7%
  44. 44 Qwen2.5-72B 90.7%
  45. 45 Voxtral Small 90.7%
  46. 46 Llama 3-70B 90.7%
  47. 47 Gemma 3 12B PT 90.7%
  48. 48 utter-project/EuroLLM-22B-Instruct-Preview 90.6%
  49. 49 speakleash/Bielik-11B-v2.3-Instruct 90.5%
  50. 50 Qwen3-Next-80B-A3B thinking 90.5%
  51. 51 Mistral 7B v0.3 90.5%
  52. 52 tiiuae/falcon-11B 90.5%
  53. 53 Llama 2-70B 90.5%
  54. 54 allenai/Olmo-3.1-32B-Instruct-SFT 90.4%
  55. 55 utter-project/EuroLLM-22B-Instruct-2512 90.4%
  56. 56 bullerwins/Magistral-Small-2506-fp8 90.3%
  57. 57 utter-project/EuroLLM-22B-Preview 90.3%
  58. 58 Gemma 2 9B 90.2%
  59. 59 deepcogito/cogito-v1-preview-llama-8B 90.2%
  60. 60 meta-llama/Llama-2-13b-hf 90.2%
  61. 61 Command R (03-2024) 90.2%
  62. 62 Granite 4.0 H Small 90.2%
  63. 63 Llama 3.1-70B 90.2%
  64. 64 GLM-4.1V-9B-Thinking 90.2%
  65. 65 Gemma 3 27B PT 90.1%
  66. 66 meta-llama/Meta-Llama-3-70B-Instruct 90.1%
  67. 67 c4ai-command-r-08-2024 90.1%
  68. 68 Qwen3-235B-A22B 90%
  69. 69 Nemotron 3 Nano 30B-A3B no-thinking 90%
  70. 70 ChocoLlama/Llama-3-ChocoLlama-8B-instruct 90%
  71. 71 meta-llama/Llama-3.1-8B-Instruct 90%
  72. 72 deepcogito/cogito-v1-preview-qwen-14B 90%
  73. 73 occiglot/occiglot-7b-eu5 89.9%
  74. 74 BSC-LT/salamandra-7b 89.9%
  75. 75 Reka Flash 3.1 89.9%
  76. 76 Yi-1.5-34B 89.9%
  77. 77 LumiOpen/Llama-Poro-2-70B-Instruct 89.8%
  78. 78 DeepSeek-R1-Distill-Llama-70B 89.8%
  79. 79 LumiOpen/Viking-33B 89.7%
  80. 80 HuggingFaceTB/SmolLM3-3B-Base 89.7%
  81. 81 Phi-4 89.7%
  82. 82 Llama 4 Scout 89.7%
  83. 83 Qwen3-8B 89.6%
  84. 84 Qwen1.5-14B-Chat 89.5%
  85. 85 allenai/OLMo-2-0325-32B 89.3%
  86. 86 ibm-granite/granite-4.0-h-tiny 89.3%
  87. 87 occiglot/occiglot-7b-eu5-instruct 89.3%
  88. 88 mistralai/Mixtral-8x7B-Instruct-v0.1 89.2%
  89. 89 152334H/miqu-1-70b-sf 89.2%
  90. 90 ibm-granite/granite-3.0-3b-a800m-base 89.2%
  91. 91 meta-llama/Meta-Llama-3-8B-Instruct 89.2%
  92. 92 ibm-granite/granite-4.0-h-micro 89.1%
  93. 93 Granite-3.0-2B-Base 89.1%
  94. 94 utter-project/EuroLLM-9B-Instruct 89%
  95. 95 Gemma 3n E2B 89%
  96. 96 CohereForAI/aya-23-8B 89%
  97. 97 mistralai/Ministral-3-8B-Reasoning-2512 89%
  98. 98 mistralai/Mistral-Nemo-Instruct-2407 89%
  99. 99 mistralai/Ministral-8B-Instruct-2410 89%
  100. 100 meta-llama/Llama-2-7b-hf 88.9%
  101. 101 typesafe/jev-1.13.0 zero-shot 88.9%
  102. 102 allenai/Olmo-3-1025-7B 88.8%
  103. 103 allenai/Olmo-3-1125-32B 88.8%
  104. 104 Llama 3.2 3B 88.8%
  105. 105 DeepSeek-R1-Distill-Qwen-32B 88.8%
  106. 106 allenai/OLMo-2-1124-13B-Instruct 88.7%
  107. 107 Gemma 3n E4B 88.6%
  108. 108 DeepSeek-R1-Distill-Qwen-14B 88.5%
  109. 109 Ministral 3 3B Reasoning 88.5%
  110. 110 utter-project/EuroLLM-22B-2512 88.5%
  111. 111 BramVanroy/GEITje-7B-ultra 88.5%
  112. 112 ibm-granite/granite-3.0-3b-a800m-instruct 88.4%
  113. 113 Qwen3-4B-Thinking-2507 88.4%
  114. 114 NousResearch/DeepHermes-3-Llama-3-8B-Preview 88.4%
  115. 115 deepseek-r1-0528-qwen3-8b 88.4%
  116. 116 Gemma 7B 88.3%
  117. 117 allenai/Llama-3.1-Tulu-3-8B-SFT 88.2%
  118. 118 Granite-3.0-8B-Instruct 88.1%
  119. 119 mistralai/Mistral-Small-3.1-24B-Base-2503 88.1%
  120. 120 Mistral 7B v0.1 88%
  121. 121 allenai/OLMo-2-1124-13B 88%
  122. 122 neuralmagic/Sparse-Llama-3.1-8B-2of4 88%
  123. 123 Granite-3.0-8B-Base 88%
  124. 124 gpt-oss-20b medium 87.9%
  125. 125 norallm/normistral-11b-warm 87.8%
  126. 126 upstage/SOLAR-10.7B-v1.0 87.8%
  127. 127 01-ai/Yi-1.5-6B 87.8%
  128. 128 Falcon3-7B-Instruct 87.7%
  129. 129 CohereForAI/aya-23-35B 87.6%
  130. 130 nvidia/mistral-nemo-minitron-8b-base 87.6%
  131. 131 swiss-ai/Apertus-8B-2509 87.5%
  132. 132 ibm-granite/granite-34b-code-base 87.5%
  133. 133 Qwen1.5-14B 87.5%
  134. 134 google/gemma-2-2b 87.5%
  135. 135 mistralai/Voxtral-Mini-3B-2507 87.4%
  136. 136 granite-4.0-micro 87.2%
  137. 137 meta-llama/Llama-3.2-3B-Instruct 87.1%
  138. 138 utter-project/EuroLLM-9B 86.9%
  139. 139 tiiuae/Falcon3-7B-Base 86.9%
  140. 140 deepcogito/cogito-v1-preview-llama-3B 86.7%
  141. 141 nvidia/mistral-nemo-minitron-8b-instruct 86.6%
  142. 142 Gemma 3 4B 86.6%
  143. 143 HuggingFaceTB/SmolLM3-3B no-thinking 86.6%
  144. 144 mistralai/Ministral-3-3B-Base-2512 86.5%
  145. 145 Gemma 2B 86.5%
  146. 146 Mistral-7B-Instruct-v0.2 86.5%
  147. 147 qwen3-4b-instruct-2507 86.3%
  148. 148 TrustLLMeu/baseline-7-8b_2-3t-tokens_llama 86.2%
  149. 149 Granite-3.0-2B-Instruct 86.1%
  150. 150 Gemma 3 4B PT 86.1%
  151. 151 BSC-LT/salamandra-7b-instruct 86.1%
  152. 152 mii-llm/nesso-4B 86%
  153. 153 allenai/OLMo-2-1124-7B-Instruct 85.9%
  154. 154 stabilityai/stablelm-2-1_6b 85.8%
  155. 155 speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551 zero-shot 85.8%
  156. 156 Starling-LM-7B-beta 85.6%
  157. 157 nvidia-nemotron-nano-9b-v2 85.6%
  158. 158 ibm-granite/granite-34b-code-instruct 85.4%
  159. 159 speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c zero-shot 85.3%
  160. 160 Command R7B 85.3%
  161. 161 Gemma 2 27B 85.3%
  162. 162 intfloat/multilingual-e5-large 85.3%
  163. 163 ibm-granite/granite-20b-code-base 85.3%
  164. 164 NorwAI/NorwAI-Mistral-7B 85.2%
  165. 165 allenai/OLMo-2-1124-7B 85.1%
  166. 166 TildeAI/TildeOpen-30b slow-tokenizer 85.1%
  167. 167 ollama_chat/gemma3:4b-it-q4_K_M 85%
  168. 168 google/rembert 85%
  169. 169 Gemma 2 2B IT 84.9%
  170. 170 AFM-4.5B 84.8%
  171. 171 baidu/ERNIE-4.5-21B-A3B-PT 84.8%
  172. 172 senseable/WestLake-7B-v2 84.7%
  173. 173 LumiOpen/Viking-13B 84.6%
  174. 174 speakleash/Bielik-11B-v2.6-Instruct 84.5%
  175. 175 meta-llama/Llama-2-13b-chat-hf 84.4%
  176. 176 Qwen3-4B no-thinking 84.3%
  177. 177 meta-llama/Llama-2-70b-chat-hf 84.2%
  178. 178 Apertus 8B Instruct 84.2%
  179. 179 TrustLLMeu/baseline-7-8b_1t-tokens_llama 84.1%
  180. 180 MaLA-LM/emma-500-llama2-7b 83.9%
  181. 181 TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49 83.8%
  182. 182 Qwen/Qwen1.5-4B 83.7%
  183. 183 Mistral-7B-Instruct-v0.1 83.7%
  184. 184 ollama_chat/hf.co/google/gemma-3-4b-it-qat-q4_0-gguf 83.2%
  185. 185 Trinity Mini 82.9%
  186. 186 arcee-ai/AFM-4.5B-Base 82.9%
  187. 187 Qwen3-1.7B no-thinking 82.9%
  188. 188 ChocoLlama/ChocoLlama-2-7B-base 82.8%
  189. 189 Qwen/Qwen2.5-Math-7B 82.7%
  190. 190 ibm-granite/granite-7b-base 82.4%
  191. 191 Qwen1.5-4B-Chat 82.3%
  192. 192 ChocoLlama/Llama-3-ChocoLlama-8B-base 81.5%
  193. 193 EuroBERT/EuroBERT-610m 81%
  194. 194 Llama 3.2 1B 80.6%
  195. 195 Qwen/Qwen2.5-1.5B-Instruct 80.1%
  196. 196 ibm-granite/granite-8b-code-base-4k 80.1%
  197. 197 jhu-clsp/mmBERT-base 80%
  198. 198 allenai/Olmo-3-7B-Instruct 79.9%
  199. 199 allenai/Olmo-3-7B-Instruct-DPO 79.8%
  200. 200 allenai/Olmo-3-7B-Think 79.6%
  201. 201 ibm-granite/granite-8b-code-instruct-4k 79.5%
  202. 202 Phi-3-mini-128k-instruct 79.4%
  203. 203 allenai/Olmo-3-7B-Instruct-SFT 79.1%
  204. 204 Nemotron 3 Nano 4B 78.7%
  205. 205 LiquidAI/LFM2-2.6B 78.7%
  206. 206 OmniJev/OneJev1.6-0.8B zero-shot 78.1%
  207. 207 setu4993/LaBSE 77.2%
  208. 208 deepseek-ai/DeepSeek-R1-Distill-Llama-8B 77%
  209. 209 intfloat/multilingual-e5-base 77%
  210. 210 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6 75.2%
  211. 211 Hunyuan-4B-Instruct 74.9%
  212. 212 ZurichNLP/unsup-simcse-xlm-roberta-base 74.9%
  213. 213 syvai/claes-preview zero-shot 74.8%
  214. 214 EuroBERT/EuroBERT-210m 74.6%
  215. 215 timpal0l/Mistral-7B-v0.1-flashback-v2 74.5%
  216. 216 DeepSeek-R1-Distill-Qwen-7B 74.4%
  217. 217 SmolLM2-1.7B 74.3%
  218. 218 SmolLM2 1.7B Instruct 74.1%
  219. 219 Gemma 3 1B 74.1%
  220. 220 ibm-granite/granite-3.0-1b-a400m-base 73.4%
  221. 221 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f 73.2%
  222. 222 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660 72.2%
  223. 223 TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894 71.4%
  224. 224 AI-Sweden-Models/ModernBERT-large 71.1%
  225. 225 common-pile/comma-v0.1-2t 71%
  226. 226 convaiinnovations/laya-multilingual zero-shot 70.8%
  227. 227 ibm-granite/granite-20b-code-instruct 70.7%
  228. 228 HPLT/gpt-13b-nordic-prerelease 70.2%
  229. 229 jaredpalmer/kev-0.5b zero-shot 70.1%
  230. 230 state-spaces/mamba-2.8b-hf 70%
  231. 231 infly/OpenCoder-8B-Instruct 69.4%
  232. 232 jhu-clsp/mmBERT-small 69.2%
  233. 233 HPLT/hplt2c_nld_checkpoints 68.8%
  234. 234 ChocoLlama/ChocoLlama-2-7B-tokentrans-base 68.5%
  235. 235 FacebookAI/xlm-roberta-large 68.3%
  236. 236 speakleash/Bielik-11B-v3-Base-20250730 68%
  237. 237 infly/OpenCoder-8B-Base 67.7%
  238. 238 Qwen/Qwen1.5-1.8B-Chat 67.1%
  239. 239 ChocoLlama/ChocoLlama-2-7B-instruct 66.9%
  240. 240 utter-project/EuroMoE-2.6B-A0.6B-Preview 66.5%
  241. 241 tencent/Hunyuan-7B-Instruct 66.4%
  242. 242 arcee-ai/Trinity-Mini-Base 65.8%
  243. 243 syvai/claes-v1 zero-shot 65.8%
  244. 244 LFM2.5-1.2B-Instruct 65.5%
  245. 245 common-pile/comma-v0.1-1t 65.1%
  246. 246 fastino/GLiNER2.5-Decide zero-shot 65%
  247. 247 ibm-granite/granite-3.0-1b-a400m-instruct 64.6%
  248. 248 Qwen3-0.6B no-thinking 64.1%
  249. 249 openGPT-X/Teuken-7B-instruct-research-v0.4 63.1%
  250. 250 yulan-team/YuLan-Mini-Instruct 62.2%
  251. 251 microsoft/mdeberta-v3-base 61.5%
  252. 252 Falcon3-3B-Instruct 61%
  253. 253 openGPT-X/Teuken-7B-instruct-commercial-v0.4 60.1%
  254. 254 sentence-transformers/stsb-xlm-r-multilingual 59.6%
  255. 255 google/gemma-2b-it 58.9%
  256. 256 LiquidAI/LFM2.5-1.2B-Base 58.1%
  257. 257 facebook/MobileLLM-R1-950M-base 58%
  258. 258 Gemma 3 1B PT 56.4%
  259. 259 Twitter/twhin-bert-base 55.1%
  260. 260 arcee-ai/Trinity-Nano-Base 54.9%
  261. 261 swiss-ai/Apertus-v1.1-1.5B zero-shot 53.2%
  262. 262 PleIAs/Pleias-3b-Preview 53.2%
  263. 263 BSC-LT/ALIA-40b 53.1%
  264. 264 meta-llama/Llama-2-7b-chat-hf 52.4%
  265. 265 Phi-4-mini-reasoning 51.7%
  266. 266 ibm-granite/granite-3b-code-base-2k 49.6%
  267. 267 google/gemma-7b-it 49.3%
  268. 268 RuterNorway/Llama-2-7b-chat-norwegian 49.2%
  269. 269 sentence-transformers/paraphrase-xlm-r-multilingual-v1 48.3%
  270. 270 PleIAs/Pleias-Nano 48.2%
  271. 271 microsoft/xlm-align-base 47.7%
  272. 272 tiiuae/Falcon3-3B-Base 46.9%
  273. 273 NorwAI/NorwAI-Mixtral-8x7B-instruct 45%
  274. 274 speakleash/Bielik-4.5B-v3.0-Instruct 44.8%
  275. 275 tencent/Hunyuan-1.8B-Instruct 44.6%
  276. 276 Phi-4-reasoning 43%
  277. 277 sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 40.3%
  278. 278 Falcon3-1B-Instruct 38.7%
  279. 279 sentence-transformers/distiluse-base-multilingual-cased-v1 38.5%
  280. 280 LFM2-1.2B 37.9%
  281. 281 PleIAs/Pleias-1.2b-Preview 37.6%
  282. 282 sentence-transformers/distiluse-base-multilingual-cased 37.2%
  283. 283 sentence-transformers/distiluse-base-multilingual-cased-v2 37.2%
  284. 284 speakleash/Bielik-1.5B-v3.0-Instruct 37.2%
  285. 285 distilbert/distilbert-base-multilingual-cased 35%
  286. 286 DeepSeek-R1-Distill-Qwen-1.5B 33.9%
  287. 287 sentence-transformers/quora-distilbert-multilingual 33.9%
  288. 288 sentence-transformers/distilbert-multilingual-nli-stsb-quora-ranking 33.8%
  289. 289 answerdotai/ModernBERT-base 33.5%
  290. 290 LFM2.5-1.2B-Thinking 33.1%
  291. 291 AI-Sweden-Models/gpt-sw3-20b 32.8%
  292. 292 Twitter/twhin-bert-large 31.2%
  293. 293 AI-Sweden-Models/gpt-sw3-20b-instruct 30.1%
  294. 294 EuropeanParliament/EUBERT 29.5%
  295. 295 Geotrend/distilbert-base-25lang-cased 29.5%
  296. 296 answerdotai/ModernBERT-large 28.3%
  297. 297 AI-Sweden-Models/roberta-large-1160k 27.5%
  298. 298 facebook/MobileLLM-R1-360M-base 27.5%
  299. 299 Qwen/Qwen1.5-1.8B 27.3%
  300. 300 arcee-ai/Trinity-Nano-Preview 27%
  301. 301 Qwen/Qwen2.5-Math-1.5B 26.7%
  302. 302 infly/OpenCoder-1.5B-Instruct 25.5%
  303. 303 cardiffnlp/twitter-xlm-roberta-base 25.4%
  304. 304 Phi-2 24.2%
  305. 305 AI-Sweden-Models/bert-large-nordic-pile-1M-steps 23.9%
  306. 306 ibm-granite/granite-3b-code-instruct-2k 23.6%
  307. 307 convaiinnovations/laya zero-shot 22.1%
  308. 308 Gemma 3 270M 21.8%
  309. 309 meta-llama/Llama-3.2-1B-Instruct 17.8%
  310. 310 Phi-4-Reasoning-plus 17.5%
  311. 311 HPLT/hplt2c_eng_checkpoints 14.9%
  312. 312 SmolLM2-360M 14.6%
  313. 313 LiquidAI/LFM2-350M 13.7%
  314. 314 dbmdz/bert-base-historic-multilingual-cased 13.6%
  315. 315 LiquidAI/LFM2-700M 13.2%
  316. 316 Qwen/Qwen1.5-0.5B-Chat 13.2%
  317. 317 Qwen/Qwen1.5-0.5B 10.9%
  318. 318 utter-project/EuroVLM-1.7B-Preview 10.3%
  319. 319 PleIAs/Pleias-Pico 9.67%
  320. 320 fresh-xlm-roberta-base 9.54%
  321. 321 PleIAs/Pleias-350m-Preview 8.96%
  322. 322 baidu/ERNIE-4.5-0.3B-PT 7.64%
  323. 323 HPLT/hplt2c_deu_checkpoints 6.87%
  324. 324 facebook/MobileLLM-R1-140M-base 6.81%
  325. 325 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview 5.61%
  326. 326 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base 4.28%
  327. 327 DiscoResearch/DiscoLM-70b 3.50%
  328. 328 AI-Sweden-Models/gpt-sw3-356m-instruct 2.62%
  329. 329 dbmdz/bert-tiny-historic-multilingual-cased 2.13%
  330. 330 SmolLM2-135M 1.78%
  331. 331 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct 0.91%
  332. 332 PleIAs/Baguettotron@refs/pr/6 no-thinking 0.86%
  333. 333 AI-Sweden-Models/gpt-sw3-126m-instruct 0.67%
  334. 334 AI-Sweden-Models/gpt-sw3-356m 0.38%
  335. 335 fresh-electra-small 0.23%
  336. 336 infly/OpenCoder-1.5B-Base 0.23%
  337. 337 facebook/MobileLLM-R1-360M 0.06%
  338. 338 ChocoLlama/ChocoLlama-2-7B-tokentrans-instruct 0%
  339. 339 facebook/MobileLLM-R1-950M 0%
  340. 340 kyutai/helium-1-2b 0%
  341. 341 openai-community/gpt2 0%
  342. 342 openai-community/gpt2-large 0%
  343. 343 openai-community/gpt2-medium 0%
  344. 344 Shekswess/trlm-135m 0%
  345. 345 facebook/MobileLLM-R1-140M -0.76%
  346. 346 HuggingFaceTB/SmolLM2-135M-Instruct -0.98%
  347. 347 PleIAs/Monad -1.28%
  348. 348 AI-Sweden-Models/gpt-sw3-126m -1.60%
  349. 349 ibm-granite/granite-7b-instruct -2.34%
  350. 350 ssmits/Falcon2-5.5B-multilingual -4.94%
  351. 351 HuggingFaceTB/SmolLM2-360M-Instruct -6.52%
  352. 352 RJuro/kanelsnegl-v0.1 -13.48%

Source: EuroEval’s leaderboard

105 models. Scroll sideways for the rest.

  1. 92.5 OLMo 2 32B Instruct
  2. 92.5 Gemma 3 27B
  3. 92.3 Llama 3.1-405B
  4. 92.1 Qwen3-Next-80B-A3B-Instruct
  5. 92.1 Llama 3.3 70B Instruct
  6. 91.9 Apertus 70B
  7. 91.9 Mistral Small 3.1
  8. 91.9 Llama 3-8B
  9. 91.8 Qwen3-14B
  10. 91.7 Aya Expanse 8B
  11. 91.7 Qwen3-32B
  12. 91.7 Llama 3.1-8B
  13. 91.5 Llama-3.1-Tulu-3-8B
  14. 91.4 Olmo 3.1 32B Instruct
  15. 91.4 Mistral Small 3.2
  16. 91.3 Aya Expanse 32B
  17. 91.3 Cogito v1 Preview 70B
  18. 91.3 Llama-3.1-Nemotron-70B-Instruct
  19. 91.2 Gemma 3 12B
  20. 91.0 Mixtral 8x7B
  21. 90.8 GLM-4.5-Air
  22. 90.8 QwQ-32B
  23. 90.7 Qwen3-30B-A3B
  24. 90.7 Reka Flash 3
  25. 90.7 Qwen2.5-72B
  26. 90.7 Voxtral Small
  27. 90.7 Llama 3-70B
  28. 90.7 Gemma 3 12B PT
  29. 90.5 Qwen3-Next-80B-A3B
  30. 90.5 Mistral 7B v0.3
  31. 90.5 Llama 2-70B
  32. 90.2 Gemma 2 9B
  33. 90.2 Command R (03-2024)
  34. 90.2 Granite 4.0 H Small
  35. 90.2 Llama 3.1-70B
  36. 90.2 GLM-4.1V-9B-Thinking
  37. 90.1 Gemma 3 27B PT
  38. 90.1 c4ai-command-r-08-2024
  39. 90.0 Qwen3-235B-A22B
  40. 90.0 Nemotron 3 Nano 30B-A3B
  41. 89.9 Reka Flash 3.1
  42. 89.9 Yi-1.5-34B
  43. 89.8 DeepSeek-R1-Distill-Llama-70B
  44. 89.7 Phi-4
  45. 89.7 Llama 4 Scout
  46. 89.6 Qwen3-8B
  47. 89.5 Qwen1.5-14B-Chat
  48. 89.1 Granite-3.0-2B-Base
  49. 89.0 Gemma 3n E2B
  50. 88.8 Llama 3.2 3B
  51. 88.8 DeepSeek-R1-Distill-Qwen-32B
  52. 88.6 Gemma 3n E4B
  53. 88.5 DeepSeek-R1-Distill-Qwen-14B
  54. 88.5 Ministral 3 3B Reasoning
  55. 88.4 Qwen3-4B-Thinking-2507
  56. 88.4 deepseek-r1-0528-qwen3-8b
  57. 88.3 Gemma 7B
  58. 88.1 Granite-3.0-8B-Instruct
  59. 88.0 Mistral 7B v0.1
  60. 88.0 Granite-3.0-8B-Base
  61. 87.9 gpt-oss-20b
  62. 87.7 Falcon3-7B-Instruct
  63. 87.5 Qwen1.5-14B
  64. 87.2 granite-4.0-micro
  65. 86.6 Gemma 3 4B
  66. 86.5 Gemma 2B
  67. 86.5 Mistral-7B-Instruct-v0.2
  68. 86.3 qwen3-4b-instruct-2507
  69. 86.1 Granite-3.0-2B-Instruct
  70. 86.1 Gemma 3 4B PT
  71. 85.6 Starling-LM-7B-beta
  72. 85.6 nvidia-nemotron-nano-9b-v2
  73. 85.3 Command R7B
  74. 85.3 Gemma 2 27B
  75. 84.9 Gemma 2 2B IT
  76. 84.8 AFM-4.5B
  77. 84.3 Qwen3-4B
  78. 84.2 Apertus 8B Instruct
  79. 83.7 Mistral-7B-Instruct-v0.1
  80. 82.9 Trinity Mini
  81. 82.9 Qwen3-1.7B
  82. 82.3 Qwen1.5-4B-Chat
  83. 80.6 Llama 3.2 1B
  84. 79.4 Phi-3-mini-128k-instruct
  85. 78.7 Nemotron 3 Nano 4B
  86. 74.9 Hunyuan-4B-Instruct
  87. 74.4 DeepSeek-R1-Distill-Qwen-7B
  88. 74.3 SmolLM2-1.7B
  89. 74.1 SmolLM2 1.7B Instruct
  90. 74.1 Gemma 3 1B
  91. 65.5 LFM2.5-1.2B-Instruct
  92. 64.1 Qwen3-0.6B
  93. 61.0 Falcon3-3B-Instruct
  94. 56.4 Gemma 3 1B PT
  95. 51.7 Phi-4-mini-reasoning
  96. 43.0 Phi-4-reasoning
  97. 38.7 Falcon3-1B-Instruct
  98. 37.9 LFM2-1.2B
  99. 33.9 DeepSeek-R1-Distill-Qwen-1.5B
  100. 33.1 LFM2.5-1.2B-Thinking
  101. 24.2 Phi-2
  102. 21.8 Gemma 3 270M
  103. 17.5 Phi-4-Reasoning-plus
  104. 14.6 SmolLM2-360M
  105. 1.78 SmolLM2-135M
  1. 1 OLMo 2 32B Instruct 92.5%
  2. 2 Gemma 3 27B 92.5%
  3. 3 Llama 3.1-405B 92.3%
  4. 4 Qwen3-Next-80B-A3B-Instruct 92.1%
  5. 5 Llama 3.3 70B Instruct 92.1%
  6. 6 Apertus 70B 91.9%
  7. 7 Mistral Small 3.1 91.9%
  8. 8 Llama 3-8B 91.9%
  9. 9 Qwen3-14B no-thinking 91.8%
  10. 10 Aya Expanse 8B 91.7%
  11. 11 Qwen3-32B no-thinking 91.7%
  12. 12 Llama 3.1-8B 91.7%
  13. 13 Llama-3.1-Tulu-3-8B 91.5%
  14. 14 Olmo 3.1 32B Instruct 91.4%
  15. 15 Mistral Small 3.2 91.4%
  16. 16 Aya Expanse 32B 91.3%
  17. 17 Cogito v1 Preview 70B 91.3%
  18. 18 Llama-3.1-Nemotron-70B-Instruct 91.3%
  19. 19 Gemma 3 12B 91.2%
  20. 20 Mixtral 8x7B 91%
  21. 21 GLM-4.5-Air thinking 90.8%
  22. 22 QwQ-32B 90.8%
  23. 23 Qwen3-30B-A3B no-thinking 90.7%
  24. 24 Reka Flash 3 90.7%
  25. 25 Qwen2.5-72B 90.7%
  26. 26 Voxtral Small 90.7%
  27. 27 Llama 3-70B 90.7%
  28. 28 Gemma 3 12B PT 90.7%
  29. 29 Qwen3-Next-80B-A3B thinking 90.5%
  30. 30 Mistral 7B v0.3 90.5%
  31. 31 Llama 2-70B 90.5%
  32. 32 Gemma 2 9B 90.2%
  33. 33 Command R (03-2024) 90.2%
  34. 34 Granite 4.0 H Small 90.2%
  35. 35 Llama 3.1-70B 90.2%
  36. 36 GLM-4.1V-9B-Thinking 90.2%
  37. 37 Gemma 3 27B PT 90.1%
  38. 38 c4ai-command-r-08-2024 90.1%
  39. 39 Qwen3-235B-A22B 90%
  40. 40 Nemotron 3 Nano 30B-A3B no-thinking 90%
  41. 41 Reka Flash 3.1 89.9%
  42. 42 Yi-1.5-34B 89.9%
  43. 43 DeepSeek-R1-Distill-Llama-70B 89.8%
  44. 44 Phi-4 89.7%
  45. 45 Llama 4 Scout 89.7%
  46. 46 Qwen3-8B 89.6%
  47. 47 Qwen1.5-14B-Chat 89.5%
  48. 48 Granite-3.0-2B-Base 89.1%
  49. 49 Gemma 3n E2B 89%
  50. 50 Llama 3.2 3B 88.8%
  51. 51 DeepSeek-R1-Distill-Qwen-32B 88.8%
  52. 52 Gemma 3n E4B 88.6%
  53. 53 DeepSeek-R1-Distill-Qwen-14B 88.5%
  54. 54 Ministral 3 3B Reasoning 88.5%
  55. 55 Qwen3-4B-Thinking-2507 88.4%
  56. 56 deepseek-r1-0528-qwen3-8b 88.4%
  57. 57 Gemma 7B 88.3%
  58. 58 Granite-3.0-8B-Instruct 88.1%
  59. 59 Mistral 7B v0.1 88%
  60. 60 Granite-3.0-8B-Base 88%
  61. 61 gpt-oss-20b medium 87.9%
  62. 62 Falcon3-7B-Instruct 87.7%
  63. 63 Qwen1.5-14B 87.5%
  64. 64 granite-4.0-micro 87.2%
  65. 65 Gemma 3 4B 86.6%
  66. 66 Gemma 2B 86.5%
  67. 67 Mistral-7B-Instruct-v0.2 86.5%
  68. 68 qwen3-4b-instruct-2507 86.3%
  69. 69 Granite-3.0-2B-Instruct 86.1%
  70. 70 Gemma 3 4B PT 86.1%
  71. 71 Starling-LM-7B-beta 85.6%
  72. 72 nvidia-nemotron-nano-9b-v2 85.6%
  73. 73 Command R7B 85.3%
  74. 74 Gemma 2 27B 85.3%
  75. 75 Gemma 2 2B IT 84.9%
  76. 76 AFM-4.5B 84.8%
  77. 77 Qwen3-4B no-thinking 84.3%
  78. 78 Apertus 8B Instruct 84.2%
  79. 79 Mistral-7B-Instruct-v0.1 83.7%
  80. 80 Trinity Mini 82.9%
  81. 81 Qwen3-1.7B no-thinking 82.9%
  82. 82 Qwen1.5-4B-Chat 82.3%
  83. 83 Llama 3.2 1B 80.6%
  84. 84 Phi-3-mini-128k-instruct 79.4%
  85. 85 Nemotron 3 Nano 4B 78.7%
  86. 86 Hunyuan-4B-Instruct 74.9%
  87. 87 DeepSeek-R1-Distill-Qwen-7B 74.4%
  88. 88 SmolLM2-1.7B 74.3%
  89. 89 SmolLM2 1.7B Instruct 74.1%
  90. 90 Gemma 3 1B 74.1%
  91. 91 LFM2.5-1.2B-Instruct 65.5%
  92. 92 Qwen3-0.6B no-thinking 64.1%
  93. 93 Falcon3-3B-Instruct 61%
  94. 94 Gemma 3 1B PT 56.4%
  95. 95 Phi-4-mini-reasoning 51.7%
  96. 96 Phi-4-reasoning 43%
  97. 97 Falcon3-1B-Instruct 38.7%
  98. 98 LFM2-1.2B 37.9%
  99. 99 DeepSeek-R1-Distill-Qwen-1.5B 33.9%
  100. 100 LFM2.5-1.2B-Thinking 33.1%
  101. 101 Phi-2 24.2%
  102. 102 Gemma 3 270M 21.8%
  103. 103 Phi-4-Reasoning-plus 17.5%
  104. 104 SmolLM2-360M 14.6%
  105. 105 SmolLM2-135M 1.78%

Source: EuroEval’s leaderboard

#EntryScore ±SettingDateLinkSource
1 Qwen/Qwen3-235B-A22B-Instruct-2507-FP8 closed weights 93% ±0.2 hf.co Official leaderboard
2 OLMo 2 32B Instruct 92.5% ±0.3 Mar 2025 hf.co Official leaderboard
3 Gemma 3 27B 92.5% ±0.3 Mar 2025 hf.co Official leaderboard
4 Llama 3.1-405B 92.3% ±0.4 Jul 2024 hf.co Official leaderboard
5 NovaSky-AI/Sky-T1-32B-Preview open weights 92.3% ±0.3 hf.co Official leaderboard
6 Qwen3-Next-80B-A3B-Instruct 92.1% ±0.3 Sep 2025 hf.co Official leaderboard
7 Llama 3.3 70B Instruct 92.1% ±0.2 Dec 2024 hf.co Official leaderboard
8 ThatsGroes/gemma-2-27b-it-FP8-Dynamic open weights 92% ±0.3 hf.co Official leaderboard
9 Apertus 70B 91.9% ±0.4 Sep 2025 hf.co Official leaderboard
10 Mistral Small 3.1 91.9% ±0.5 Mar 2025 hf.co Official leaderboard
11 Llama 3-8B 91.9% ±0.1 Apr 2024 hf.co Official leaderboard
12 Qwen3-14B 91.8% ±0.3 no-thinking Apr 2025 hf.co Official leaderboard
13 Aya Expanse 8B 91.7% ±0.3 Oct 2024 hf.co Official leaderboard
14 Qwen3-32B 91.7% ±0.3 no-thinking Apr 2025 hf.co Official leaderboard
15 Llama 3.1-8B 91.7% ±0.1 Jul 2024 hf.co Official leaderboard
16 allenai/Olmo-3.1-32B-Instruct-DPO open weights 91.6% ±0.4 hf.co Official leaderboard
17 Llama-3.1-Tulu-3-8B 91.5% ±0.3 Nov 2024 hf.co Official leaderboard
18 allenai/Llama-3.1-Tulu-3-8B-DPO open weights 91.5% ±0.3 hf.co Official leaderboard
19 Olmo 3.1 32B Instruct 91.4% ±0.4 Dec 2025 hf.co Official leaderboard
20 Qwen/Qwen3-235B-A22B-Thinking-2507-FP8 closed weights 91.4% ±0.3 hf.co Official leaderboard
21 Mistral Small 3.2 91.4% ±0.4 Jun 2025 hf.co Official leaderboard
22 01-ai/Yi-1.5-9B open weights 91.4% ±0.4 hf.co Official leaderboard
23 ai21labs/Jamba-v0.1 open weights 91.4% ±0.7 hf.co Official leaderboard
24 allenai/Llama-3.1-Tulu-3.1-8B open weights 91.4% ±0.3 hf.co Official leaderboard
25 Aya Expanse 32B 91.3% ±0.3 Oct 2024 hf.co Official leaderboard
26 Cogito v1 Preview 70B 91.3% ±0.4 Apr 2025 hf.co Official leaderboard
27 google/gemma-2-9b-it open weights 91.3% ±0.4 hf.co Official leaderboard
28 Llama-3.1-Nemotron-70B-Instruct 91.3% ±0.4 Oct 2024 hf.co Official leaderboard
29 google/gemma-2-27b-it open weights 91.3% ±0.4 hf.co Official leaderboard
30 LumiOpen/Llama-Poro-2-70B-SFT open weights 91.2% ±0.4 hf.co Official leaderboard
31 Gemma 3 12B 91.2% ±0.5 Mar 2025 hf.co Official leaderboard
32 mistralai/Magistral-Small-2509 open weights 91.1% ±0.4 hf.co Official leaderboard
33 deepcogito/cogito-v2-preview-llama-70B open weights 91.1% ±0.4 hf.co Official leaderboard
34 meta-llama/Llama-3.1-405B-Instruct-FP8 open weights 91.1% ±0.4 hf.co Official leaderboard
35 Mixtral 8x7B 91% ±0.6 Dec 2023 hf.co Official leaderboard
36 deepcogito/cogito-v1-preview-qwen-32B open weights 91% ±0.4 hf.co Official leaderboard
37 meta-llama/Llama-4-Scout-17B-16E open weights 90.9% ±0.7 hf.co Official leaderboard
38 GLM-4.5-Air 90.8% ±0.4 thinking Jul 2025 hf.co Official leaderboard
39 QwQ-32B 90.8% ±0.3 Mar 2025 hf.co Official leaderboard
40 meta-llama/Llama-3.1-70B-Instruct open weights 90.8% ±0.4 hf.co Official leaderboard
41 Qwen3-30B-A3B 90.7% ±0.3 no-thinking Apr 2025 hf.co Official leaderboard
42 mistralai/Mistral-Small-24B-Instruct-2501 open weights 90.7% ±0.4 hf.co Official leaderboard
43 Reka Flash 3 90.7% ±0.3 Mar 2025 hf.co Official leaderboard
44 Qwen2.5-72B 90.7% ±0.4 Sep 2024 hf.co Official leaderboard
45 Voxtral Small 90.7% ±0.4 Jul 2025 hf.co Official leaderboard
46 Llama 3-70B 90.7% ±0.5 Apr 2024 hf.co Official leaderboard
47 Gemma 3 12B PT 90.7% ±0.9 Mar 2025 hf.co Official leaderboard
48 utter-project/EuroLLM-22B-Instruct-Preview open weights 90.6% ±0.3 hf.co Official leaderboard
49 speakleash/Bielik-11B-v2.3-Instruct open weights 90.5% ±0.3 hf.co Official leaderboard
50 Qwen3-Next-80B-A3B 90.5% ±0.3 thinking Sep 2025 hf.co Official leaderboard
51 Mistral 7B v0.3 90.5% ±0.5 Sep 2023 hf.co Official leaderboard
52 tiiuae/falcon-11B open weights 90.5% ±0.3 hf.co Official leaderboard
53 Llama 2-70B 90.5% ±1.4 Jul 2023 hf.co Official leaderboard
54 allenai/Olmo-3.1-32B-Instruct-SFT open weights 90.4% ±0.3 hf.co Official leaderboard
55 utter-project/EuroLLM-22B-Instruct-2512 open weights 90.4% ±0.2 hf.co Official leaderboard
56 bullerwins/Magistral-Small-2506-fp8 open weights 90.3% ±0.6 hf.co Official leaderboard
57 utter-project/EuroLLM-22B-Preview open weights 90.3% ±0.4 hf.co Official leaderboard
58 Gemma 2 9B 90.2% ±0.6 Jun 2024 hf.co Official leaderboard
59 deepcogito/cogito-v1-preview-llama-8B open weights 90.2% ±0.3 hf.co Official leaderboard
60 meta-llama/Llama-2-13b-hf open weights 90.2% ±0.8 hf.co Official leaderboard
61 Command R (03-2024) 90.2% ±0.4 Mar 2024 hf.co Official leaderboard
62 Granite 4.0 H Small 90.2% ±0.3 Oct 2025 hf.co Official leaderboard
63 Llama 3.1-70B 90.2% ±0.5 Jul 2024 hf.co Official leaderboard
64 GLM-4.1V-9B-Thinking 90.2% ±0.3 Jul 2025 hf.co Official leaderboard
65 Gemma 3 27B PT 90.1% ±0.7 Mar 2025 hf.co Official leaderboard
66 meta-llama/Meta-Llama-3-70B-Instruct open weights 90.1% ±0.5 hf.co Official leaderboard
67 c4ai-command-r-08-2024 90.1% ±0.4 Aug 2024 hf.co Official leaderboard
68 Qwen3-235B-A22B 90% ±0.4 Apr 2025 hf.co Official leaderboard
69 Nemotron 3 Nano 30B-A3B 90% ±0.3 no-thinking Dec 2025 hf.co Official leaderboard
70 ChocoLlama/Llama-3-ChocoLlama-8B-instruct open weights 90% ±0.4 hf.co Official leaderboard
71 meta-llama/Llama-3.1-8B-Instruct open weights 90% ±0.3 hf.co Official leaderboard
72 deepcogito/cogito-v1-preview-qwen-14B open weights 90% ±0.5 hf.co Official leaderboard
73 occiglot/occiglot-7b-eu5 open weights 89.9% ±0.7 hf.co Official leaderboard
74 BSC-LT/salamandra-7b open weights 89.9% ±0.5 hf.co Official leaderboard
75 Reka Flash 3.1 89.9% ±0.4 May 2025 hf.co Official leaderboard
76 Yi-1.5-34B 89.9% ±0.6 May 2024 hf.co Official leaderboard
77 LumiOpen/Llama-Poro-2-70B-Instruct open weights 89.8% ±0.5 hf.co Official leaderboard
78 DeepSeek-R1-Distill-Llama-70B 89.8% ±0.4 Jan 2025 hf.co Official leaderboard
79 LumiOpen/Viking-33B open weights 89.7% ±0.5 hf.co Official leaderboard
80 HuggingFaceTB/SmolLM3-3B-Base open weights 89.7% ±0.7 hf.co Official leaderboard
81 Phi-4 89.7% ±0.3 Dec 2024 hf.co Official leaderboard
82 Llama 4 Scout 89.7% ±0.6 Apr 2025 hf.co Official leaderboard
83 Qwen3-8B 89.6% ±0.4 Apr 2025 hf.co Official leaderboard
84 Qwen1.5-14B-Chat 89.5% ±0.5 Feb 2024 hf.co Official leaderboard
85 allenai/OLMo-2-0325-32B open weights 89.3% ±0.4 hf.co Official leaderboard
86 ibm-granite/granite-4.0-h-tiny open weights 89.3% ±0.6 hf.co Official leaderboard
87 occiglot/occiglot-7b-eu5-instruct open weights 89.3% ±0.6 hf.co Official leaderboard
88 mistralai/Mixtral-8x7B-Instruct-v0.1 open weights 89.2% ±0.5 hf.co Official leaderboard
89 152334H/miqu-1-70b-sf open weights 89.2% ±0.3 hf.co Official leaderboard
90 ibm-granite/granite-3.0-3b-a800m-base open weights 89.2% ±0.7 hf.co Official leaderboard
91 meta-llama/Meta-Llama-3-8B-Instruct open weights 89.2% ±0.7 hf.co Official leaderboard
92 ibm-granite/granite-4.0-h-micro open weights 89.1% ±0.4 hf.co Official leaderboard
93 Granite-3.0-2B-Base 89.1% ±0.6 Oct 2024 hf.co Official leaderboard
94 utter-project/EuroLLM-9B-Instruct open weights 89% ±0.5 hf.co Official leaderboard
95 Gemma 3n E2B 89% ±0.4 Jun 2025 hf.co Official leaderboard
96 CohereForAI/aya-23-8B open weights 89% ±0.6 hf.co Official leaderboard
97 mistralai/Ministral-3-8B-Reasoning-2512 open weights 89% ±0.5 hf.co Official leaderboard
98 mistralai/Mistral-Nemo-Instruct-2407 open weights 89% ±0.6 hf.co Official leaderboard
99 mistralai/Ministral-8B-Instruct-2410 open weights 89% ±0.4 hf.co Official leaderboard
100 meta-llama/Llama-2-7b-hf open weights 88.9% ±0.8 hf.co Official leaderboard
101 typesafe/jev-1.13.0 closed weights 88.9% ±0.3 zero-shot docs.typesafe.ai Official leaderboard
102 allenai/Olmo-3-1025-7B open weights 88.8% ±0.4 hf.co Official leaderboard
103 allenai/Olmo-3-1125-32B open weights 88.8% ±0.6 hf.co Official leaderboard
104 Llama 3.2 3B 88.8% ±0.4 Sep 2024 hf.co Official leaderboard
105 DeepSeek-R1-Distill-Qwen-32B 88.8% ±0.5 Jan 2025 hf.co Official leaderboard
106 allenai/OLMo-2-1124-13B-Instruct open weights 88.7% ±0.4 hf.co Official leaderboard
107 Gemma 3n E4B 88.6% ±0.5 Jun 2025 hf.co Official leaderboard
108 DeepSeek-R1-Distill-Qwen-14B 88.5% ±0.3 Jan 2025 hf.co Official leaderboard
109 Ministral 3 3B Reasoning 88.5% ±0.3 Dec 2025 hf.co Official leaderboard
110 utter-project/EuroLLM-22B-2512 open weights 88.5% ±2.1 hf.co Official leaderboard
111 BramVanroy/GEITje-7B-ultra open weights 88.5% ±0.6 hf.co Official leaderboard
112 ibm-granite/granite-3.0-3b-a800m-instruct open weights 88.4% ±0.4 hf.co Official leaderboard
113 Qwen3-4B-Thinking-2507 88.4% ±0.3 Aug 2025 hf.co Official leaderboard
114 NousResearch/DeepHermes-3-Llama-3-8B-Preview open weights 88.4% ±0.5 hf.co Official leaderboard
115 deepseek-r1-0528-qwen3-8b 88.4% ±0.3 May 2025 hf.co Official leaderboard
116 Gemma 7B 88.3% ±0.9 Feb 2024 hf.co Official leaderboard
117 allenai/Llama-3.1-Tulu-3-8B-SFT open weights 88.2% ±0.4 hf.co Official leaderboard
118 Granite-3.0-8B-Instruct 88.1% ±0.6 Oct 2024 hf.co Official leaderboard
119 mistralai/Mistral-Small-3.1-24B-Base-2503 open weights 88.1% ±2.7 hf.co Official leaderboard
120 Mistral 7B v0.1 88% ±0.8 Sep 2023 hf.co Official leaderboard
121 allenai/OLMo-2-1124-13B open weights 88% ±0.5 hf.co Official leaderboard
122 neuralmagic/Sparse-Llama-3.1-8B-2of4 open weights 88% ±0.3 hf.co Official leaderboard
123 Granite-3.0-8B-Base 88% ±0.9 Oct 2024 hf.co Official leaderboard
124 gpt-oss-20b 87.9% ±0.3 medium Aug 2025 hf.co Official leaderboard
125 norallm/normistral-11b-warm open weights 87.8% ±0.5 hf.co Official leaderboard
126 upstage/SOLAR-10.7B-v1.0 open weights 87.8% ±0.6 hf.co Official leaderboard
127 01-ai/Yi-1.5-6B open weights 87.8% ±0.7 hf.co Official leaderboard
128 Falcon3-7B-Instruct 87.7% ±0.5 Dec 2024 hf.co Official leaderboard
129 CohereForAI/aya-23-35B open weights 87.6% ±0.5 hf.co Official leaderboard
130 nvidia/mistral-nemo-minitron-8b-base open weights 87.6% ±0.9 hf.co Official leaderboard
131 swiss-ai/Apertus-8B-2509 open weights 87.5% ±0.8 hf.co Official leaderboard
132 ibm-granite/granite-34b-code-base open weights 87.5% ±0.4 hf.co Official leaderboard
133 Qwen1.5-14B 87.5% ±0.4 Feb 2024 hf.co Official leaderboard
134 google/gemma-2-2b open weights 87.5% ±0.9 hf.co Official leaderboard
135 mistralai/Voxtral-Mini-3B-2507 open weights 87.4% ±0.4 hf.co Official leaderboard
136 granite-4.0-micro 87.2% ±0.7 Oct 2025 hf.co Official leaderboard
137 meta-llama/Llama-3.2-3B-Instruct open weights 87.1% ±0.4 hf.co Official leaderboard
138 utter-project/EuroLLM-9B open weights 86.9% ±1.9 hf.co Official leaderboard
139 tiiuae/Falcon3-7B-Base open weights 86.9% ±0.3 hf.co Official leaderboard
140 deepcogito/cogito-v1-preview-llama-3B open weights 86.7% ±0.8 hf.co Official leaderboard
141 nvidia/mistral-nemo-minitron-8b-instruct open weights 86.6% ±0.6 hf.co Official leaderboard
142 Gemma 3 4B 86.6% ±0.7 Mar 2025 hf.co Official leaderboard
143 HuggingFaceTB/SmolLM3-3B open weights 86.6% ±0.7 no-thinking hf.co Official leaderboard
144 mistralai/Ministral-3-3B-Base-2512 open weights 86.5% ±1 hf.co Official leaderboard
145 Gemma 2B 86.5% ±1 Feb 2024 hf.co Official leaderboard
146 Mistral-7B-Instruct-v0.2 86.5% ±0.6 Dec 2023 hf.co Official leaderboard
147 qwen3-4b-instruct-2507 86.3% ±0.7 Aug 2025 hf.co Official leaderboard
148 TrustLLMeu/baseline-7-8b_2-3t-tokens_llama open weights 86.2% ±1.1 hf.co Official leaderboard
149 Granite-3.0-2B-Instruct 86.1% ±0.3 Oct 2024 hf.co Official leaderboard
150 Gemma 3 4B PT 86.1% ±1 Mar 2025 hf.co Official leaderboard
151 BSC-LT/salamandra-7b-instruct open weights 86.1% ±0.5 hf.co Official leaderboard
152 mii-llm/nesso-4B open weights 86% ±0.9 hf.co Official leaderboard
153 allenai/OLMo-2-1124-7B-Instruct open weights 85.9% ±0.5 hf.co Official leaderboard
154 stabilityai/stablelm-2-1_6b open weights 85.8% ±0.6 hf.co Official leaderboard
155 speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551 open weights 85.8% ±0.2 zero-shot hf.co Official leaderboard
156 Starling-LM-7B-beta 85.6% ±0.6 Mar 2024 hf.co Official leaderboard
157 nvidia-nemotron-nano-9b-v2 85.6% ±0.3 Aug 2025 hf.co Official leaderboard
158 ibm-granite/granite-34b-code-instruct open weights 85.4% ±0.4 hf.co Official leaderboard
159 speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c open weights 85.3% ±0.3 zero-shot hf.co Official leaderboard
160 Command R7B 85.3% ±0.7 Dec 2024 hf.co Official leaderboard
161 Gemma 2 27B 85.3% ±5.9 Jun 2024 hf.co Official leaderboard
162 intfloat/multilingual-e5-large open weights 85.3% ±0.5 hf.co Official leaderboard
163 ibm-granite/granite-20b-code-base open weights 85.3% ±0.7 hf.co Official leaderboard
164 NorwAI/NorwAI-Mistral-7B open weights 85.2% ±0.9 hf.co Official leaderboard
165 allenai/OLMo-2-1124-7B open weights 85.1% ±0.8 hf.co Official leaderboard
166 TildeAI/TildeOpen-30b open weights 85.1% ±0.9 slow-tokenizer hf.co Official leaderboard
167 ollama_chat/gemma3:4b-it-q4_K_M open weights 85% ±0.6 ollama.com Official leaderboard
168 google/rembert open weights 85% ±0.5 hf.co Official leaderboard
169 Gemma 2 2B IT 84.9% ±0.8 Jul 2024 hf.co Official leaderboard
170 AFM-4.5B 84.8% ±0.6 Jul 2025 hf.co Official leaderboard
171 baidu/ERNIE-4.5-21B-A3B-PT open weights 84.8% ±1 hf.co Official leaderboard
172 senseable/WestLake-7B-v2 open weights 84.7% ±0.7 hf.co Official leaderboard
173 LumiOpen/Viking-13B open weights 84.6% ±1 hf.co Official leaderboard
174 speakleash/Bielik-11B-v2.6-Instruct open weights 84.5% ±0.8 hf.co Official leaderboard
175 meta-llama/Llama-2-13b-chat-hf open weights 84.4% ±1.6 hf.co Official leaderboard
176 Qwen3-4B 84.3% ±0.7 no-thinking Apr 2025 hf.co Official leaderboard
177 meta-llama/Llama-2-70b-chat-hf open weights 84.2% ±2.2 hf.co Official leaderboard
178 Apertus 8B Instruct 84.2% ±1 Sep 2025 hf.co Official leaderboard
179 TrustLLMeu/baseline-7-8b_1t-tokens_llama open weights 84.1% ±1.1 hf.co Official leaderboard
180 MaLA-LM/emma-500-llama2-7b open weights 83.9% ±0.8 hf.co Official leaderboard
181 TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49 open weights 83.8% ±0.5 hf.co Official leaderboard
182 Qwen/Qwen1.5-4B open weights 83.7% ±0.5 hf.co Official leaderboard
183 Mistral-7B-Instruct-v0.1 83.7% ±0.9 Sep 2023 hf.co Official leaderboard
184 ollama_chat/hf.co/google/gemma-3-4b-it-qat-q4_0-gguf open weights 83.2% ±0.7 ollama.com Official leaderboard
185 Trinity Mini 82.9% ±0.5 Dec 2025 hf.co Official leaderboard
186 arcee-ai/AFM-4.5B-Base open weights 82.9% ±0.6 hf.co Official leaderboard
187 Qwen3-1.7B 82.9% ±0.7 no-thinking Apr 2025 hf.co Official leaderboard
188 ChocoLlama/ChocoLlama-2-7B-base open weights 82.8% ±1.2 hf.co Official leaderboard
189 Qwen/Qwen2.5-Math-7B open weights 82.7% ±0.6 hf.co Official leaderboard
190 ibm-granite/granite-7b-base open weights 82.4% ±1 hf.co Official leaderboard
191 Qwen1.5-4B-Chat 82.3% ±0.6 Feb 2024 hf.co Official leaderboard
192 ChocoLlama/Llama-3-ChocoLlama-8B-base open weights 81.5% ±2.1 hf.co Official leaderboard
193 EuroBERT/EuroBERT-610m open weights 81% ±1 hf.co Official leaderboard
194 Llama 3.2 1B 80.6% ±1.5 Sep 2024 hf.co Official leaderboard
195 Qwen/Qwen2.5-1.5B-Instruct open weights 80.1% ±0.9 hf.co Official leaderboard
196 ibm-granite/granite-8b-code-base-4k open weights 80.1% ±0.6 hf.co Official leaderboard
197 jhu-clsp/mmBERT-base open weights 80% ±0.8 hf.co Official leaderboard
198 allenai/Olmo-3-7B-Instruct open weights 79.9% ±0.6 hf.co Official leaderboard
199 allenai/Olmo-3-7B-Instruct-DPO open weights 79.8% ±0.7 hf.co Official leaderboard
200 allenai/Olmo-3-7B-Think open weights 79.6% ±0.9 hf.co Official leaderboard
201 ibm-granite/granite-8b-code-instruct-4k open weights 79.5% ±1.2 hf.co Official leaderboard
202 Phi-3-mini-128k-instruct 79.4% ±0.6 Apr 2024 hf.co Official leaderboard
203 allenai/Olmo-3-7B-Instruct-SFT open weights 79.1% ±1.2 hf.co Official leaderboard
204 Nemotron 3 Nano 4B 78.7% ±1.3 Mar 2026 hf.co Official leaderboard
205 LiquidAI/LFM2-2.6B open weights 78.7% ±0.6 hf.co Official leaderboard
206 OmniJev/OneJev1.6-0.8B open weights 78.1% ±0.3 zero-shot hf.co Official leaderboard
207 setu4993/LaBSE open weights 77.2% ±0.6 hf.co Official leaderboard
208 deepseek-ai/DeepSeek-R1-Distill-Llama-8B open weights 77% ±1.4 hf.co Official leaderboard
209 intfloat/multilingual-e5-base open weights 77% ±1.1 hf.co Official leaderboard
210 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6 open weights 75.2% ±1.9 hf.co Official leaderboard
211 Hunyuan-4B-Instruct 74.9% ±0.9 Jul 2025 hf.co Official leaderboard
212 ZurichNLP/unsup-simcse-xlm-roberta-base open weights 74.9% ±0.5 hf.co Official leaderboard
213 syvai/claes-preview open weights 74.8% ±0.3 zero-shot hf.co Official leaderboard
214 EuroBERT/EuroBERT-210m open weights 74.6% ±1.7 hf.co Official leaderboard
215 timpal0l/Mistral-7B-v0.1-flashback-v2 open weights 74.5% ±1.6 hf.co Official leaderboard
216 DeepSeek-R1-Distill-Qwen-7B 74.4% ±0.9 Jan 2025 hf.co Official leaderboard
217 SmolLM2-1.7B 74.3% ±1.5 Oct 2024 hf.co Official leaderboard
218 SmolLM2 1.7B Instruct 74.1% ±1 Oct 2024 hf.co Official leaderboard
219 Gemma 3 1B 74.1% ±1.8 Mar 2025 hf.co Official leaderboard
220 ibm-granite/granite-3.0-1b-a400m-base open weights 73.4% ±0.9 hf.co Official leaderboard
221 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f open weights 73.2% ±3.1 hf.co Official leaderboard
222 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660 open weights 72.2% ±3.1 hf.co Official leaderboard
223 TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894 open weights 71.4% ±2.7 hf.co Official leaderboard
224 AI-Sweden-Models/ModernBERT-large open weights 71.1% ±3.2 hf.co Official leaderboard
225 common-pile/comma-v0.1-2t open weights 71% ±1.1 hf.co Official leaderboard
226 convaiinnovations/laya-multilingual open weights 70.8% ±0.4 zero-shot hf.co Official leaderboard
227 ibm-granite/granite-20b-code-instruct open weights 70.7% ±2.9 hf.co Official leaderboard
228 HPLT/gpt-13b-nordic-prerelease open weights 70.2% ±1.6 hf.co Official leaderboard
229 jaredpalmer/kev-0.5b open weights 70.1% ±0.4 zero-shot hf.co Official leaderboard
230 state-spaces/mamba-2.8b-hf open weights 70% ±2.3 hf.co Official leaderboard
231 infly/OpenCoder-8B-Instruct open weights 69.4% ±1.2 hf.co Official leaderboard
232 jhu-clsp/mmBERT-small open weights 69.2% ±1.9 hf.co Official leaderboard
233 HPLT/hplt2c_nld_checkpoints open weights 68.8% ±2.3 hf.co Official leaderboard
234 ChocoLlama/ChocoLlama-2-7B-tokentrans-base open weights 68.5% ±6.6 hf.co Official leaderboard
235 FacebookAI/xlm-roberta-large open weights 68.3% ±11.4 hf.co Official leaderboard
236 speakleash/Bielik-11B-v3-Base-20250730 open weights 68% ±6.1 hf.co Official leaderboard
237 infly/OpenCoder-8B-Base open weights 67.7% ±0.8 hf.co Official leaderboard
238 Qwen/Qwen1.5-1.8B-Chat open weights 67.1% ±1.5 hf.co Official leaderboard
239 ChocoLlama/ChocoLlama-2-7B-instruct open weights 66.9% ±3.8 hf.co Official leaderboard
240 utter-project/EuroMoE-2.6B-A0.6B-Preview open weights 66.5% ±2.3 hf.co Official leaderboard
241 tencent/Hunyuan-7B-Instruct open weights 66.4% ±0.9 hf.co Official leaderboard
242 arcee-ai/Trinity-Mini-Base open weights 65.8% ±3.4 hf.co Official leaderboard
243 syvai/claes-v1 open weights 65.8% ±0.4 zero-shot hf.co Official leaderboard
244 LFM2.5-1.2B-Instruct 65.5% ±2 Jan 2026 hf.co Official leaderboard
245 common-pile/comma-v0.1-1t open weights 65.1% ±0.6 hf.co Official leaderboard
246 fastino/GLiNER2.5-Decide open weights 65% ±0.4 zero-shot hf.co Official leaderboard
247 ibm-granite/granite-3.0-1b-a400m-instruct open weights 64.6% ±1.4 hf.co Official leaderboard
248 Qwen3-0.6B 64.1% ±2.3 no-thinking Apr 2025 hf.co Official leaderboard
249 openGPT-X/Teuken-7B-instruct-research-v0.4 open weights 63.1% ±2.3 hf.co Official leaderboard
250 yulan-team/YuLan-Mini-Instruct open weights 62.2% ±2.6 hf.co Official leaderboard
251 microsoft/mdeberta-v3-base open weights 61.5% ±10.3 hf.co Official leaderboard
252 Falcon3-3B-Instruct 61% ±0.7 Dec 2024 hf.co Official leaderboard
253 openGPT-X/Teuken-7B-instruct-commercial-v0.4 open weights 60.1% ±2 hf.co Official leaderboard
254 sentence-transformers/stsb-xlm-r-multilingual open weights 59.6% ±0.6 hf.co Official leaderboard
255 google/gemma-2b-it open weights 58.9% ±2.3 hf.co Official leaderboard
256 LiquidAI/LFM2.5-1.2B-Base open weights 58.1% ±3 hf.co Official leaderboard
257 facebook/MobileLLM-R1-950M-base open weights 58% ±1.1 hf.co Official leaderboard
258 Gemma 3 1B PT 56.4% ±4.8 Mar 2025 hf.co Official leaderboard
259 Twitter/twhin-bert-base open weights 55.1% ±3.1 hf.co Official leaderboard
260 arcee-ai/Trinity-Nano-Base open weights 54.9% ±7.3 hf.co Official leaderboard
261 swiss-ai/Apertus-v1.1-1.5B open weights 53.2% ±0.3 zero-shot hf.co Official leaderboard
262 PleIAs/Pleias-3b-Preview open weights 53.2% ±2.7 hf.co Official leaderboard
263 BSC-LT/ALIA-40b open weights 53.1% ±8.2 hf.co Official leaderboard
264 meta-llama/Llama-2-7b-chat-hf open weights 52.4% ±3.5 hf.co Official leaderboard
265 Phi-4-mini-reasoning 51.7% ±2.8 Apr 2025 hf.co Official leaderboard
266 ibm-granite/granite-3b-code-base-2k open weights 49.6% ±3.2 hf.co Official leaderboard
267 google/gemma-7b-it open weights 49.3% ±5.6 hf.co Official leaderboard
268 RuterNorway/Llama-2-7b-chat-norwegian open weights 49.2% ±4.9 hf.co Official leaderboard
269 sentence-transformers/paraphrase-xlm-r-multilingual-v1 open weights 48.3% ±6.4 hf.co Official leaderboard
270 PleIAs/Pleias-Nano open weights 48.2% ±5.7 hf.co Official leaderboard
271 microsoft/xlm-align-base open weights 47.7% ±12.1 hf.co Official leaderboard
272 tiiuae/Falcon3-3B-Base open weights 46.9% ±1.2 hf.co Official leaderboard
273 NorwAI/NorwAI-Mixtral-8x7B-instruct open weights 45% ±9.7 hf.co Official leaderboard
274 speakleash/Bielik-4.5B-v3.0-Instruct open weights 44.8% ±14.9 hf.co Official leaderboard
275 tencent/Hunyuan-1.8B-Instruct open weights 44.6% ±1.5 hf.co Official leaderboard
276 Phi-4-reasoning 43% ±1.8 Apr 2025 hf.co Official leaderboard
277 sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 open weights 40.3% ±5.9 hf.co Official leaderboard
278 Falcon3-1B-Instruct 38.7% ±0.9 Dec 2024 hf.co Official leaderboard
279 sentence-transformers/distiluse-base-multilingual-cased-v1 open weights 38.5% ±4 hf.co Official leaderboard
280 LFM2-1.2B 37.9% ±3 Jul 2025 hf.co Official leaderboard
281 PleIAs/Pleias-1.2b-Preview open weights 37.6% ±4.4 hf.co Official leaderboard
282 sentence-transformers/distiluse-base-multilingual-cased open weights 37.2% ±2.4 hf.co Official leaderboard
283 sentence-transformers/distiluse-base-multilingual-cased-v2 open weights 37.2% ±2.4 hf.co Official leaderboard
284 speakleash/Bielik-1.5B-v3.0-Instruct open weights 37.2% ±12.4 hf.co Official leaderboard
285 distilbert/distilbert-base-multilingual-cased open weights 35% ±5.1 hf.co Official leaderboard
286 DeepSeek-R1-Distill-Qwen-1.5B 33.9% ±1.4 Jan 2025 hf.co Official leaderboard
287 sentence-transformers/quora-distilbert-multilingual open weights 33.9% ±3.6 hf.co Official leaderboard
288 sentence-transformers/distilbert-multilingual-nli-stsb-quora-ranking open weights 33.8% ±3.6 hf.co Official leaderboard
289 answerdotai/ModernBERT-base open weights 33.5% ±2.7 hf.co Official leaderboard
290 LFM2.5-1.2B-Thinking 33.1% ±3.3 Jan 2026 hf.co Official leaderboard
291 AI-Sweden-Models/gpt-sw3-20b open weights 32.8% ±3.1 hf.co Official leaderboard
292 Twitter/twhin-bert-large open weights 31.2% ±9.9 hf.co Official leaderboard
293 AI-Sweden-Models/gpt-sw3-20b-instruct open weights 30.1% ±5.4 hf.co Official leaderboard
294 EuropeanParliament/EUBERT open weights 29.5% ±1.2 hf.co Official leaderboard
295 Geotrend/distilbert-base-25lang-cased open weights 29.5% ±5.4 hf.co Official leaderboard
296 answerdotai/ModernBERT-large open weights 28.3% ±4.3 hf.co Official leaderboard
297 AI-Sweden-Models/roberta-large-1160k open weights 27.5% ±4.5 hf.co Official leaderboard
298 facebook/MobileLLM-R1-360M-base open weights 27.5% ±1.9 hf.co Official leaderboard
299 Qwen/Qwen1.5-1.8B open weights 27.3% ±4 hf.co Official leaderboard
300 arcee-ai/Trinity-Nano-Preview open weights 27% ±4.9 hf.co Official leaderboard
301 Qwen/Qwen2.5-Math-1.5B open weights 26.7% ±3.4 hf.co Official leaderboard
302 infly/OpenCoder-1.5B-Instruct open weights 25.5% ±3.2 hf.co Official leaderboard
303 cardiffnlp/twitter-xlm-roberta-base open weights 25.4% ±7.7 hf.co Official leaderboard
304 Phi-2 24.2% ±2.7 Dec 2023 hf.co Official leaderboard
305 AI-Sweden-Models/bert-large-nordic-pile-1M-steps open weights 23.9% ±2.6 hf.co Official leaderboard
306 ibm-granite/granite-3b-code-instruct-2k open weights 23.6% ±4.8 hf.co Official leaderboard
307 convaiinnovations/laya open weights 22.1% ±0.8 zero-shot hf.co Official leaderboard
308 Gemma 3 270M 21.8% ±4.4 Aug 2025 hf.co Official leaderboard
309 meta-llama/Llama-3.2-1B-Instruct open weights 17.8% ±3.5 hf.co Official leaderboard
310 Phi-4-Reasoning-plus 17.5% ±4.4 Apr 2025 hf.co Official leaderboard
311 HPLT/hplt2c_eng_checkpoints open weights 14.9% ±2.8 hf.co Official leaderboard
312 SmolLM2-360M 14.6% ±3.4 Oct 2024 hf.co Official leaderboard
313 LiquidAI/LFM2-350M open weights 13.7% ±3.3 hf.co Official leaderboard
314 dbmdz/bert-base-historic-multilingual-cased open weights 13.6% ±2.1 hf.co Official leaderboard
315 LiquidAI/LFM2-700M open weights 13.2% ±1.5 hf.co Official leaderboard
316 Qwen/Qwen1.5-0.5B-Chat open weights 13.2% ±3.6 hf.co Official leaderboard
317 Qwen/Qwen1.5-0.5B open weights 10.9% ±2.3 hf.co Official leaderboard
318 utter-project/EuroVLM-1.7B-Preview open weights 10.3% ±1.7 hf.co Official leaderboard
319 PleIAs/Pleias-Pico open weights 9.67% ±1.9 hf.co Official leaderboard
320 fresh-xlm-roberta-base open weights 9.54% ±4 Official leaderboard
321 PleIAs/Pleias-350m-Preview open weights 8.96% ±3.2 hf.co Official leaderboard
322 baidu/ERNIE-4.5-0.3B-PT open weights 7.64% ±1.7 hf.co Official leaderboard
323 HPLT/hplt2c_deu_checkpoints open weights 6.87% ±1.8 hf.co Official leaderboard
324 facebook/MobileLLM-R1-140M-base open weights 6.81% ±1.2 hf.co Official leaderboard
325 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview open weights 5.61% ±1.4 hf.co Official leaderboard
326 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base open weights 4.28% ±1 hf.co Official leaderboard
327 DiscoResearch/DiscoLM-70b open weights 3.50% ±0.5 hf.co Official leaderboard
328 AI-Sweden-Models/gpt-sw3-356m-instruct open weights 2.62% ±1.5 hf.co Official leaderboard
329 dbmdz/bert-tiny-historic-multilingual-cased open weights 2.13% ±1 hf.co Official leaderboard
330 SmolLM2-135M 1.78% ±1.4 Oct 2024 hf.co Official leaderboard
331 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct open weights 0.91% ±0.5 hf.co Official leaderboard
332 PleIAs/Baguettotron@refs/pr/6 open weights 0.86% ±1.1 no-thinking hf.co Official leaderboard
333 AI-Sweden-Models/gpt-sw3-126m-instruct open weights 0.67% ±0.8 hf.co Official leaderboard
334 AI-Sweden-Models/gpt-sw3-356m open weights 0.38% ±3.2 hf.co Official leaderboard
335 fresh-electra-small open weights 0.23% ±0.2 Official leaderboard
336 infly/OpenCoder-1.5B-Base open weights 0.23% ±0.8 hf.co Official leaderboard
337 facebook/MobileLLM-R1-360M open weights 0.06% ±0.4 hf.co Official leaderboard
338 ChocoLlama/ChocoLlama-2-7B-tokentrans-instruct open weights 0% hf.co Official leaderboard
339 facebook/MobileLLM-R1-950M open weights 0% hf.co Official leaderboard
340 kyutai/helium-1-2b open weights 0% hf.co Official leaderboard
341 openai-community/gpt2 open weights 0% hf.co Official leaderboard
342 openai-community/gpt2-large open weights 0% hf.co Official leaderboard
343 openai-community/gpt2-medium open weights 0% hf.co Official leaderboard
344 Shekswess/trlm-135m open weights 0% hf.co Official leaderboard
345 facebook/MobileLLM-R1-140M open weights -0.76% ±0.6 hf.co Official leaderboard
346 HuggingFaceTB/SmolLM2-135M-Instruct open weights -0.98% ±0.9 hf.co Official leaderboard
347 PleIAs/Monad open weights -1.28% ±0.9 hf.co Official leaderboard
348 AI-Sweden-Models/gpt-sw3-126m open weights -1.60% ±1.9 hf.co Official leaderboard
349 ibm-granite/granite-7b-instruct open weights -2.34% ±1.8 hf.co Official leaderboard
350 ssmits/Falcon2-5.5B-multilingual open weights -4.94% ±1.3 hf.co Official leaderboard
351 HuggingFaceTB/SmolLM2-360M-Instruct open weights -6.52% ±2.4 hf.co Official leaderboard
352 RJuro/kanelsnegl-v0.1 open weights -13.48% ±2 hf.co Official leaderboard
#ModelScore ±SettingDateLinkSource
1 OLMo 2 32B Instruct 92.5% ±0.3 Mar 2025 hf.co Official leaderboard
2 Gemma 3 27B 92.5% ±0.3 Mar 2025 hf.co Official leaderboard
3 Llama 3.1-405B 92.3% ±0.4 Jul 2024 hf.co Official leaderboard
4 Qwen3-Next-80B-A3B-Instruct 92.1% ±0.3 Sep 2025 hf.co Official leaderboard
5 Llama 3.3 70B Instruct 92.1% ±0.2 Dec 2024 hf.co Official leaderboard
6 Apertus 70B 91.9% ±0.4 Sep 2025 hf.co Official leaderboard
7 Mistral Small 3.1 91.9% ±0.5 Mar 2025 hf.co Official leaderboard
8 Llama 3-8B 91.9% ±0.1 Apr 2024 hf.co Official leaderboard
9 Qwen3-14B 91.8% ±0.3 no-thinking Apr 2025 hf.co Official leaderboard
10 Aya Expanse 8B 91.7% ±0.3 Oct 2024 hf.co Official leaderboard
11 Qwen3-32B 91.7% ±0.3 no-thinking Apr 2025 hf.co Official leaderboard
12 Llama 3.1-8B 91.7% ±0.1 Jul 2024 hf.co Official leaderboard
13 Llama-3.1-Tulu-3-8B 91.5% ±0.3 Nov 2024 hf.co Official leaderboard
14 Olmo 3.1 32B Instruct 91.4% ±0.4 Dec 2025 hf.co Official leaderboard
15 Mistral Small 3.2 91.4% ±0.4 Jun 2025 hf.co Official leaderboard
16 Aya Expanse 32B 91.3% ±0.3 Oct 2024 hf.co Official leaderboard
17 Cogito v1 Preview 70B 91.3% ±0.4 Apr 2025 hf.co Official leaderboard
18 Llama-3.1-Nemotron-70B-Instruct 91.3% ±0.4 Oct 2024 hf.co Official leaderboard
19 Gemma 3 12B 91.2% ±0.5 Mar 2025 hf.co Official leaderboard
20 Mixtral 8x7B 91% ±0.6 Dec 2023 hf.co Official leaderboard
21 GLM-4.5-Air 90.8% ±0.4 thinking Jul 2025 hf.co Official leaderboard
22 QwQ-32B 90.8% ±0.3 Mar 2025 hf.co Official leaderboard
23 Qwen3-30B-A3B 90.7% ±0.3 no-thinking Apr 2025 hf.co Official leaderboard
24 Reka Flash 3 90.7% ±0.3 Mar 2025 hf.co Official leaderboard
25 Qwen2.5-72B 90.7% ±0.4 Sep 2024 hf.co Official leaderboard
26 Voxtral Small 90.7% ±0.4 Jul 2025 hf.co Official leaderboard
27 Llama 3-70B 90.7% ±0.5 Apr 2024 hf.co Official leaderboard
28 Gemma 3 12B PT 90.7% ±0.9 Mar 2025 hf.co Official leaderboard
29 Qwen3-Next-80B-A3B 90.5% ±0.3 thinking Sep 2025 hf.co Official leaderboard
30 Mistral 7B v0.3 90.5% ±0.5 Sep 2023 hf.co Official leaderboard
31 Llama 2-70B 90.5% ±1.4 Jul 2023 hf.co Official leaderboard
32 Gemma 2 9B 90.2% ±0.6 Jun 2024 hf.co Official leaderboard
33 Command R (03-2024) 90.2% ±0.4 Mar 2024 hf.co Official leaderboard
34 Granite 4.0 H Small 90.2% ±0.3 Oct 2025 hf.co Official leaderboard
35 Llama 3.1-70B 90.2% ±0.5 Jul 2024 hf.co Official leaderboard
36 GLM-4.1V-9B-Thinking 90.2% ±0.3 Jul 2025 hf.co Official leaderboard
37 Gemma 3 27B PT 90.1% ±0.7 Mar 2025 hf.co Official leaderboard
38 c4ai-command-r-08-2024 90.1% ±0.4 Aug 2024 hf.co Official leaderboard
39 Qwen3-235B-A22B 90% ±0.4 Apr 2025 hf.co Official leaderboard
40 Nemotron 3 Nano 30B-A3B 90% ±0.3 no-thinking Dec 2025 hf.co Official leaderboard
41 Reka Flash 3.1 89.9% ±0.4 May 2025 hf.co Official leaderboard
42 Yi-1.5-34B 89.9% ±0.6 May 2024 hf.co Official leaderboard
43 DeepSeek-R1-Distill-Llama-70B 89.8% ±0.4 Jan 2025 hf.co Official leaderboard
44 Phi-4 89.7% ±0.3 Dec 2024 hf.co Official leaderboard
45 Llama 4 Scout 89.7% ±0.6 Apr 2025 hf.co Official leaderboard
46 Qwen3-8B 89.6% ±0.4 Apr 2025 hf.co Official leaderboard
47 Qwen1.5-14B-Chat 89.5% ±0.5 Feb 2024 hf.co Official leaderboard
48 Granite-3.0-2B-Base 89.1% ±0.6 Oct 2024 hf.co Official leaderboard
49 Gemma 3n E2B 89% ±0.4 Jun 2025 hf.co Official leaderboard
50 Llama 3.2 3B 88.8% ±0.4 Sep 2024 hf.co Official leaderboard
51 DeepSeek-R1-Distill-Qwen-32B 88.8% ±0.5 Jan 2025 hf.co Official leaderboard
52 Gemma 3n E4B 88.6% ±0.5 Jun 2025 hf.co Official leaderboard
53 DeepSeek-R1-Distill-Qwen-14B 88.5% ±0.3 Jan 2025 hf.co Official leaderboard
54 Ministral 3 3B Reasoning 88.5% ±0.3 Dec 2025 hf.co Official leaderboard
55 Qwen3-4B-Thinking-2507 88.4% ±0.3 Aug 2025 hf.co Official leaderboard
56 deepseek-r1-0528-qwen3-8b 88.4% ±0.3 May 2025 hf.co Official leaderboard
57 Gemma 7B 88.3% ±0.9 Feb 2024 hf.co Official leaderboard
58 Granite-3.0-8B-Instruct 88.1% ±0.6 Oct 2024 hf.co Official leaderboard
59 Mistral 7B v0.1 88% ±0.8 Sep 2023 hf.co Official leaderboard
60 Granite-3.0-8B-Base 88% ±0.9 Oct 2024 hf.co Official leaderboard
61 gpt-oss-20b 87.9% ±0.3 medium Aug 2025 hf.co Official leaderboard
62 Falcon3-7B-Instruct 87.7% ±0.5 Dec 2024 hf.co Official leaderboard
63 Qwen1.5-14B 87.5% ±0.4 Feb 2024 hf.co Official leaderboard
64 granite-4.0-micro 87.2% ±0.7 Oct 2025 hf.co Official leaderboard
65 Gemma 3 4B 86.6% ±0.7 Mar 2025 hf.co Official leaderboard
66 Gemma 2B 86.5% ±1 Feb 2024 hf.co Official leaderboard
67 Mistral-7B-Instruct-v0.2 86.5% ±0.6 Dec 2023 hf.co Official leaderboard
68 qwen3-4b-instruct-2507 86.3% ±0.7 Aug 2025 hf.co Official leaderboard
69 Granite-3.0-2B-Instruct 86.1% ±0.3 Oct 2024 hf.co Official leaderboard
70 Gemma 3 4B PT 86.1% ±1 Mar 2025 hf.co Official leaderboard
71 Starling-LM-7B-beta 85.6% ±0.6 Mar 2024 hf.co Official leaderboard
72 nvidia-nemotron-nano-9b-v2 85.6% ±0.3 Aug 2025 hf.co Official leaderboard
73 Command R7B 85.3% ±0.7 Dec 2024 hf.co Official leaderboard
74 Gemma 2 27B 85.3% ±5.9 Jun 2024 hf.co Official leaderboard
75 Gemma 2 2B IT 84.9% ±0.8 Jul 2024 hf.co Official leaderboard
76 AFM-4.5B 84.8% ±0.6 Jul 2025 hf.co Official leaderboard
77 Qwen3-4B 84.3% ±0.7 no-thinking Apr 2025 hf.co Official leaderboard
78 Apertus 8B Instruct 84.2% ±1 Sep 2025 hf.co Official leaderboard
79 Mistral-7B-Instruct-v0.1 83.7% ±0.9 Sep 2023 hf.co Official leaderboard
80 Trinity Mini 82.9% ±0.5 Dec 2025 hf.co Official leaderboard
81 Qwen3-1.7B 82.9% ±0.7 no-thinking Apr 2025 hf.co Official leaderboard
82 Qwen1.5-4B-Chat 82.3% ±0.6 Feb 2024 hf.co Official leaderboard
83 Llama 3.2 1B 80.6% ±1.5 Sep 2024 hf.co Official leaderboard
84 Phi-3-mini-128k-instruct 79.4% ±0.6 Apr 2024 hf.co Official leaderboard
85 Nemotron 3 Nano 4B 78.7% ±1.3 Mar 2026 hf.co Official leaderboard
86 Hunyuan-4B-Instruct 74.9% ±0.9 Jul 2025 hf.co Official leaderboard
87 DeepSeek-R1-Distill-Qwen-7B 74.4% ±0.9 Jan 2025 hf.co Official leaderboard
88 SmolLM2-1.7B 74.3% ±1.5 Oct 2024 hf.co Official leaderboard
89 SmolLM2 1.7B Instruct 74.1% ±1 Oct 2024 hf.co Official leaderboard
90 Gemma 3 1B 74.1% ±1.8 Mar 2025 hf.co Official leaderboard
91 LFM2.5-1.2B-Instruct 65.5% ±2 Jan 2026 hf.co Official leaderboard
92 Qwen3-0.6B 64.1% ±2.3 no-thinking Apr 2025 hf.co Official leaderboard
93 Falcon3-3B-Instruct 61% ±0.7 Dec 2024 hf.co Official leaderboard
94 Gemma 3 1B PT 56.4% ±4.8 Mar 2025 hf.co Official leaderboard
95 Phi-4-mini-reasoning 51.7% ±2.8 Apr 2025 hf.co Official leaderboard
96 Phi-4-reasoning 43% ±1.8 Apr 2025 hf.co Official leaderboard
97 Falcon3-1B-Instruct 38.7% ±0.9 Dec 2024 hf.co Official leaderboard
98 LFM2-1.2B 37.9% ±3 Jul 2025 hf.co Official leaderboard
99 DeepSeek-R1-Distill-Qwen-1.5B 33.9% ±1.4 Jan 2025 hf.co Official leaderboard
100 LFM2.5-1.2B-Thinking 33.1% ±3.3 Jan 2026 hf.co Official leaderboard
101 Phi-2 24.2% ±2.7 Dec 2023 hf.co Official leaderboard
102 Gemma 3 270M 21.8% ±4.4 Aug 2025 hf.co Official leaderboard
103 Phi-4-Reasoning-plus 17.5% ±4.4 Apr 2025 hf.co Official leaderboard
104 SmolLM2-360M 14.6% ±3.4 Oct 2024 hf.co Official leaderboard
105 SmolLM2-135M 1.78% ±1.4 Oct 2024 hf.co 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
Qwen3-14B
no-thinking
91.8%
GLM-4.5-Air
no-thinking
90.8%
Qwen3-8B
no-thinking
89.6%
gpt-oss-20b
low
87.9%
nvidia-nemotron-nano-9b-v2
no-thinking
85.6%
Qwen3-4B
no-thinking
84.3%
Qwen3-1.7B
no-thinking
82.9%
Qwen3-0.6B
no-thinking
64.1%

Source: EuroEval’s leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
OLMo 2 32B Instruct 92.5%
Gemma 3 27B 92.5%
Llama 3.1-405B 92.3%
Qwen3-Next-80B-A3B-Instruct 92.1%
Llama 3.3 70B Instruct 92.1%
Apertus 70B 91.9%
Mistral Small 3.1 91.9%
Llama 3-8B 91.9%
Qwen3-14B no-thinking 91.8%
Aya Expanse 8B 91.7%
Qwen3-32B no-thinking 91.7%
Llama 3.1-8B 91.7%
Llama-3.1-Tulu-3-8B 91.5%
Olmo 3.1 32B Instruct 91.4%
Mistral Small 3.2 91.4%
Aya Expanse 32B 91.3%
Cogito v1 Preview 70B 91.3%
Llama-3.1-Nemotron-70B-Instruct 91.3%
Gemma 3 12B 91.2%
Mixtral 8x7B 91%
GLM-4.5-Air thinking 90.8%
QwQ-32B 90.8%
Qwen3-30B-A3B no-thinking 90.7%
Reka Flash 3 90.7%
Qwen2.5-72B 90.7%
Voxtral Small 90.7%
Llama 3-70B 90.7%
Gemma 3 12B PT 90.7%
Qwen3-Next-80B-A3B thinking 90.5%
Mistral 7B v0.3 90.5%
Llama 2-70B 90.5%
Gemma 2 9B 90.2%
Qwen3-14B 90.2%
Command R (03-2024) 90.2%
Granite 4.0 H Small 90.2%
Llama 3.1-70B 90.2%
GLM-4.1V-9B-Thinking 90.2%
Gemma 3 27B PT 90.1%
c4ai-command-r-08-2024 90.1%
Qwen3-235B-A22B 90%
Nemotron 3 Nano 30B-A3B no-thinking 90%
Reka Flash 3.1 89.9%
Yi-1.5-34B 89.9%
DeepSeek-R1-Distill-Llama-70B 89.8%
Apertus 70B 89.7%
Phi-4 89.7%
Llama 4 Scout 89.7%
Qwen3-8B 89.6%
Qwen1.5-14B-Chat 89.5%
Granite-3.0-2B-Base 89.1%
Gemma 3n E2B 89%
Llama 3.2 3B 88.8%
DeepSeek-R1-Distill-Qwen-32B 88.8%
Gemma 3n E4B 88.6%
DeepSeek-R1-Distill-Qwen-14B 88.5%
Ministral 3 3B Reasoning 88.5%
Qwen3-4B-Thinking-2507 88.4%
deepseek-r1-0528-qwen3-8b 88.4%
Gemma 7B 88.3%
Gemma 3n E4B 88.1%
Granite-3.0-8B-Instruct 88.1%
Mistral 7B v0.1 88%
Granite-3.0-8B-Base 88%
gpt-oss-20b medium 87.9%
Falcon3-7B-Instruct 87.7%
Qwen1.5-14B 87.5%
GLM-4.5-Air no-thinking 87.4%
gpt-oss-20b low 87.2%
granite-4.0-micro 87.2%
gpt-oss-20b high 86.9%
Gemma 3 4B 86.6%
Qwen3-8B no-thinking 86.6%
Gemma 2B 86.5%
Mistral-7B-Instruct-v0.2 86.5%
qwen3-4b-instruct-2507 86.3%
Granite-3.0-2B-Instruct 86.1%
Gemma 3 4B PT 86.1%
Starling-LM-7B-beta 85.6%
nvidia-nemotron-nano-9b-v2 85.6%
Command R7B 85.3%
Gemma 2 27B 85.3%
Gemma 2 2B IT 84.9%
AFM-4.5B 84.8%
Qwen3-4B no-thinking 84.3%
Qwen3-4B 84.3%
Apertus 8B Instruct 84.2%
Mistral-7B-Instruct-v0.1 83.7%
Trinity Mini 82.9%
Qwen3-1.7B no-thinking 82.9%
Qwen1.5-4B-Chat 82.3%
Qwen3-1.7B 81.5%
Llama 3.2 1B 80.6%
Phi-3-mini-128k-instruct 79.4%
Nemotron 3 Nano 4B 78.7%
Hunyuan-4B-Instruct 74.9%
DeepSeek-R1-Distill-Qwen-7B 74.4%
SmolLM2-1.7B 74.3%
SmolLM2 1.7B Instruct 74.1%
Gemma 3 1B 74.1%
nvidia-nemotron-nano-9b-v2 no-thinking 71.7%
LFM2.5-1.2B-Instruct 65.5%
Gemma 3 1B 64.5%
Qwen3-0.6B no-thinking 64.1%
Falcon3-3B-Instruct 61%
Qwen3-0.6B 58.2%
Gemma 3 1B PT 56.4%
Phi-4-mini-reasoning 51.7%
Phi-4-reasoning 43%
Falcon3-1B-Instruct 38.7%
LFM2-1.2B 37.9%
DeepSeek-R1-Distill-Qwen-1.5B 33.9%
LFM2.5-1.2B-Thinking 33.1%
Phi-2 24.2%
Gemma 3 270M 21.8%
Phi-4-Reasoning-plus 17.5%
SmolLM2-360M 14.6%
Gemma 3 270M 9.83%
SmolLM2-135M 1.78%

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
Qwen3-14B
no-thinking
91.8%
GLM-4.5-Air
no-thinking
90.8%
Qwen3-8B
no-thinking
89.6%
gpt-oss-20b
low
87.9%
nvidia-nemotron-nano-9b-v2
no-thinking
85.6%
Qwen3-4B
no-thinking
84.3%
Qwen3-1.7B
no-thinking
82.9%
Qwen3-0.6B
no-thinking
64.1%

Source: EuroEval’s leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
OLMo 2 32B Instruct 92.5%
Gemma 3 27B 92.5%
Llama 3.1-405B 92.3%
Qwen3-Next-80B-A3B-Instruct 92.1%
Llama 3.3 70B Instruct 92.1%
Apertus 70B 91.9%
Mistral Small 3.1 91.9%
Llama 3-8B 91.9%
Qwen3-14B no-thinking 91.8%
Aya Expanse 8B 91.7%
Qwen3-32B no-thinking 91.7%
Llama 3.1-8B 91.7%
Llama-3.1-Tulu-3-8B 91.5%
Olmo 3.1 32B Instruct 91.4%
Mistral Small 3.2 91.4%
Aya Expanse 32B 91.3%
Cogito v1 Preview 70B 91.3%
Llama-3.1-Nemotron-70B-Instruct 91.3%
Gemma 3 12B 91.2%
Mixtral 8x7B 91%
GLM-4.5-Air thinking 90.8%
QwQ-32B 90.8%
Qwen3-30B-A3B no-thinking 90.7%
Reka Flash 3 90.7%
Qwen2.5-72B 90.7%
Voxtral Small 90.7%
Llama 3-70B 90.7%
Gemma 3 12B PT 90.7%
Qwen3-Next-80B-A3B thinking 90.5%
Mistral 7B v0.3 90.5%
Llama 2-70B 90.5%
Gemma 2 9B 90.2%
Qwen3-14B 90.2%
Command R (03-2024) 90.2%
Granite 4.0 H Small 90.2%
Llama 3.1-70B 90.2%
GLM-4.1V-9B-Thinking 90.2%
Gemma 3 27B PT 90.1%
c4ai-command-r-08-2024 90.1%
Qwen3-235B-A22B 90%
Nemotron 3 Nano 30B-A3B no-thinking 90%
Reka Flash 3.1 89.9%
Yi-1.5-34B 89.9%
DeepSeek-R1-Distill-Llama-70B 89.8%
Apertus 70B 89.7%
Phi-4 89.7%
Llama 4 Scout 89.7%
Qwen3-8B 89.6%
Qwen1.5-14B-Chat 89.5%
Granite-3.0-2B-Base 89.1%
Gemma 3n E2B 89%
Llama 3.2 3B 88.8%
DeepSeek-R1-Distill-Qwen-32B 88.8%
Gemma 3n E4B 88.6%
DeepSeek-R1-Distill-Qwen-14B 88.5%
Ministral 3 3B Reasoning 88.5%
Qwen3-4B-Thinking-2507 88.4%
deepseek-r1-0528-qwen3-8b 88.4%
Gemma 7B 88.3%
Gemma 3n E4B 88.1%
Granite-3.0-8B-Instruct 88.1%
Mistral 7B v0.1 88%
Granite-3.0-8B-Base 88%
gpt-oss-20b medium 87.9%
Falcon3-7B-Instruct 87.7%
Qwen1.5-14B 87.5%
GLM-4.5-Air no-thinking 87.4%
gpt-oss-20b low 87.2%
granite-4.0-micro 87.2%
gpt-oss-20b high 86.9%
Gemma 3 4B 86.6%
Qwen3-8B no-thinking 86.6%
Gemma 2B 86.5%
Mistral-7B-Instruct-v0.2 86.5%
qwen3-4b-instruct-2507 86.3%
Granite-3.0-2B-Instruct 86.1%
Gemma 3 4B PT 86.1%
Starling-LM-7B-beta 85.6%
nvidia-nemotron-nano-9b-v2 85.6%
Command R7B 85.3%
Gemma 2 27B 85.3%
Gemma 2 2B IT 84.9%
AFM-4.5B 84.8%
Qwen3-4B no-thinking 84.3%
Qwen3-4B 84.3%
Apertus 8B Instruct 84.2%
Mistral-7B-Instruct-v0.1 83.7%
Trinity Mini 82.9%
Qwen3-1.7B no-thinking 82.9%
Qwen1.5-4B-Chat 82.3%
Qwen3-1.7B 81.5%
Llama 3.2 1B 80.6%
Phi-3-mini-128k-instruct 79.4%
Nemotron 3 Nano 4B 78.7%
Hunyuan-4B-Instruct 74.9%
DeepSeek-R1-Distill-Qwen-7B 74.4%
SmolLM2-1.7B 74.3%
SmolLM2 1.7B Instruct 74.1%
Gemma 3 1B 74.1%
nvidia-nemotron-nano-9b-v2 no-thinking 71.7%
LFM2.5-1.2B-Instruct 65.5%
Gemma 3 1B 64.5%
Qwen3-0.6B no-thinking 64.1%
Falcon3-3B-Instruct 61%
Qwen3-0.6B 58.2%
Gemma 3 1B PT 56.4%
Phi-4-mini-reasoning 51.7%
Phi-4-reasoning 43%
Falcon3-1B-Instruct 38.7%
LFM2-1.2B 37.9%
DeepSeek-R1-Distill-Qwen-1.5B 33.9%
LFM2.5-1.2B-Thinking 33.1%
Phi-2 24.2%
Gemma 3 270M 21.8%
Phi-4-Reasoning-plus 17.5%
SmolLM2-360M 14.6%
Gemma 3 270M 9.83%
SmolLM2-135M 1.78%

7 models have held number one on EuroEval’s leaderboard since 2023

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

Meta Mistral AI Google DeepMind Ai2
EuroEval leaderboard Passed 85% in Jul 2023, by Llama 2-70B

Source: EuroEval’s leaderboard

Llama 2-70B on the official leaderboard passed 85% in July 2023

That was 45 months after its release. All the scores we have, one dot each, so a model tested at several settings appears more than once.

Plot with other benchmarks
On EuroEval’s leaderboard Best so far
0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) OLMo 2 32B Instruct Llama 3.1-405B Mixtral 8x7B 0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) OLMo 2 32B Instruct Mistral 7B v0.3

Source: EuroEval’s 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 8 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 2 pass 3 notes
  • Pass Independent runs Scores from the official leaderboard (352 entries).
  • Pass Models scored 105 models we track, plus 247 other entries on its leaderboard.
  • Note Newest score The newest came 8 months ago, in March 2026, for Nemotron 3 Nano 4B. That’s expected for a saturated test.
  • Note Frontier models None of the top 10 has a score, which is fine for a retired test. The strongest tested is Qwen3-235B-A22B (#131).
  • Note Headroom Saturated. The best is 92.5%, past 85% since July 2023.
Leaderboard import 3 pass
  • Pass Importer We read the board on EuroEval: 118 rows for models we track and 247 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