Benchmark Almanac

FoQA

Mistral Small 3.2 79.9% leads since June 2025
Not in model cards

Models answer questions about passages from Faroese Wikipedia. A native speaker checked and corrected automatically generated answers.

EuroEval reports character-level F1 and exact match.

Released
February 2025
Kind
Reading in many languages
Level
School
Scored
Checked automatically
Answers
Short answer
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
Qwen3-235B-A22B #131 on the capability index
Last model added
March 2026
Scores refreshed
November 20th, 2026
336
Entries on EuroEval’s leaderboard, 91 models we track

Mistral Small 3.2 leads with 79.9% among models released by March 2026 Mistral Small 3.2 leads with 79.9% among models released by March 2026

Best score per entry from EuroEval’s leaderboard, read on November 20th, 2026. EuroEval reports character-level F1 on the FoQA 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. 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:

336 entries. Scroll sideways for the rest.

  1. 79.9 Mistral Small 3.2
  2. 79.7 meta-llama/Llama-3.1-405B-Instruct-FP8
  3. 79.6 mistralai/Magistral-Small-2509
  4. 79.5 Mistral Small 3.1
  5. 79.3 Llama 3.1-70B
  6. 78.6 google/gemma-2-9b-it
  7. 77.8 LumiOpen/Llama-Poro-2-70B-SFT
  8. 77.5 mistralai/Mistral-Small-3.1-24B-Base-2503
  9. 77.4 Cogito v1 Preview 70B
  10. 77.3 Gemma 3 12B PT
  11. 77.2 Gemma 3 27B PT
  12. 76.8 deepcogito/cogito-v2-preview-llama-70B
  13. 76.7 bullerwins/Magistral-Small-2506-fp8
  14. 76.7 Apertus 70B
  15. 76.5 deepcogito/cogito-v1-preview-qwen-32B
  16. 76.4 speakleash/Bielik-11B-v3-Base-20250730
  17. 76.2 deepcogito/cogito-v1-preview-qwen-14B
  18. 75.9 TildeAI/TildeOpen-30b
  19. 75.8 Qwen3-14B
  20. 75.7 Qwen/Qwen3-235B-A22B-Instruct-2507-FP8
  21. 75.4 Voxtral Small
  22. 75.3 Gemma 2 27B
  23. 74.2 Gemma 3n E4B
  24. 74.2 NovaSky-AI/Sky-T1-32B-Preview
  25. 73.9 Qwen3-32B
  26. 73.6 mistralai/Ministral-3-8B-Reasoning-2512
  27. 73.6 Gemma 2 9B
  28. 73.6 Qwen3-Next-80B-A3B-Instruct
  29. 73.4 Apertus 8B Instruct
  30. 73.3 Llama 4 Scout
  31. 73.1 Qwen/Qwen3-235B-A22B-Thinking-2507-FP8
  32. 73.0 LumiOpen/Llama-Poro-2-70B-Instruct
  33. 72.9 Qwen3-235B-A22B
  34. 72.7 Llama 3.1-405B
  35. 72.7 Gemma 3 27B
  36. 72.5 swiss-ai/Apertus-8B-2509
  37. 72.5 allenai/OLMo-2-0325-32B
  38. 72.5 allenai/Olmo-3-1125-32B
  39. 72.3 norallm/normistral-11b-long
  40. 72.2 Gemma 7B
  41. 72.1 utter-project/EuroLLM-22B-2512
  42. 71.8 Llama 3.1-8B
  43. 71.7 allenai/Llama-3.1-Tulu-3-8B-DPO
  44. 71.5 Granite 4.0 H Small
  45. 71.1 google/gemma-2-27b-it
  46. 71.0 Reka Flash 3
  47. 70.9 Aya Expanse 32B
  48. 70.7 mistralai/Mistral-Small-24B-Instruct-2501
  49. 70.3 mistralai/Mistral-Nemo-Instruct-2407
  50. 70.3 Llama-3.1-Tulu-3-8B
  51. 70.3 Gemma 3 12B
  52. 70.3 Gemma 3 4B PT
  53. 70.2 ThatsGroes/gemma-2-27b-it-FP8-Dynamic
  54. 70.1 allenai/Llama-3.1-Tulu-3-8B-SFT
  55. 70.0 Llama 3-8B
  56. 69.7 meta-llama/Llama-3.1-8B-Instruct
  57. 69.0 Gemma 3n E2B
  58. 69.0 allenai/Llama-3.1-Tulu-3.1-8B
  59. 68.8 norallm/normistral-11b-warm
  60. 68.7 Qwen3-8B
  61. 68.6 mistralai/Ministral-3-3B-Base-2512
  62. 68.5 ollama_chat/gemma3:27b-it-fp16
  63. 68.1 BSC-LT/ALIA-40b
  64. 67.8 QwQ-32B
  65. 67.7 ai21labs/Jamba-v0.1
  66. 67.5 syvai/llama3-da-base
  67. 67.1 utter-project/EuroLLM-22B-Preview
  68. 67.1 Llama 2-70B
  69. 66.9 Yi-1.5-34B
  70. 66.7 Qwen3-30B-A3B
  71. 66.6 Falcon3-10B-Instruct
  72. 66.3 ollama_chat/gemma3:4b-it-q4_K_M
  73. 65.9 qwen3-4b-instruct-2507
  74. 65.7 NousResearch/DeepHermes-3-Llama-3-8B-Preview
  75. 65.5 Qwen3-Next-80B-A3B
  76. 65.1 upstage/SOLAR-10.7B-v1.0
  77. 65.0 baidu/ERNIE-4.5-21B-A3B-PT
  78. 64.9 CohereForAI/aya-23-35B
  79. 64.8 Qwen3-4B-Thinking-2507
  80. 64.2 Command R7B
  81. 64.0 utter-project/EuroLLM-9B
  82. 63.6 ollama_chat/hf.co/google/gemma-3-4b-it-qat-q4_0-gguf
  83. 63.2 tiiuae/Falcon3-10B-Base
  84. 63.1 mistralai/Voxtral-Mini-3B-2507
  85. 62.6 01-ai/Yi-1.5-9B
  86. 62.6 TrustLLMeu/baseline-7-8b_2-3t-tokens_llama
  87. 62.5 neuralmagic/Sparse-Llama-3.1-8B-2of4
  88. 62.3 allenai/OLMo-2-1124-13B-Instruct
  89. 62.2 mii-llm/nesso-4B
  90. 62.1 utter-project/EuroLLM-9B-Instruct
  91. 62.0 allenai/OLMo-2-1124-13B
  92. 61.9 Reka Flash 3.1
  93. 61.6 gpt-oss-20b
  94. 61.5 Phi-4-reasoning
  95. 61.4 ibm-granite/granite-34b-code-base
  96. 61.3 TrustLLMeu/baseline-7-8b_1t-tokens_llama
  97. 61.2 Qwen3-4B
  98. 61.1 Mistral 7B v0.3
  99. 61.0 Falcon3-7B-Instruct
  100. 61.0 Nemotron 3 Nano 30B-A3B
  101. 60.4 LumiOpen/Viking-33B
  102. 60.3 senseable/WestLake-7B-v2
  103. 60.1 Qwen1.5-14B-Chat
  104. 59.5 bineric/lynx-instruct-30b
  105. 58.4 Olmo 3.1 32B Instruct
  106. 58.3 utter-project/EuroLLM-22B-Instruct-2512
  107. 58.1 nvidia-nemotron-nano-9b-v2
  108. 57.4 Granite-3.0-8B-Base
  109. 57.1 Mistral 7B v0.1
  110. 56.3 meta-llama/Llama-3.2-3B-Instruct
  111. 56.2 utter-project/EuroLLM-22B-Instruct-Preview
  112. 55.9 allenai/OLMo-2-1124-7B
  113. 55.7 ibm-granite/granite-34b-code-instruct
  114. 55.6 allenai/Olmo-3.1-32B-Instruct-SFT
  115. 55.5 meta-llama/Llama-2-13b-hf
  116. 55.2 allenai/OLMo-2-1124-7B-Instruct
  117. 55.2 Granite-3.0-8B-Instruct
  118. 55.0 allenai/Olmo-3.1-32B-Instruct-DPO
  119. 54.5 norallm/normistral-11b-thinking
  120. 54.4 tiiuae/Falcon3-7B-Base
  121. 54.2 ibm-granite/granite-4.0-h-micro
  122. 54.2 ibm-granite/granite-20b-code-base
  123. 53.7 google/gemma-2-2b
  124. 53.1 TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49
  125. 53.1 allenai/Olmo-3-1025-7B
  126. 52.9 deepseek-r1-0528-qwen3-8b
  127. 52.8 speakleash/Bielik-11B-v2.3-Instruct
  128. 52.5 speakleash/Bielik-4.5B-v3.0-Instruct
  129. 52.1 LumiOpen/Viking-13B
  130. 51.2 DeepSeek-R1-Distill-Llama-70B
  131. 51.2 occiglot/occiglot-7b-eu5
  132. 51.2 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6
  133. 50.3 granite-4.0-micro
  134. 49.9 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f
  135. 49.6 Gemma 2 2B IT
  136. 49.6 HuggingFaceTB/SmolLM3-3B-Base
  137. 49.4 ibm-granite/granite-8b-code-base-4k
  138. 49.1 152334H/miqu-1-70b-sf
  139. 48.4 TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894
  140. 48.2 ltg/norbert3-large
  141. 47.0 AI-Sweden-Models/ModernBERT-large
  142. 46.8 infly/OpenCoder-8B-Instruct
  143. 45.9 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660
  144. 45.6 Ministral 3 3B Reasoning
  145. 44.9 HuggingFaceTB/SmolLM3-3B
  146. 44.9 ibm-granite/granite-8b-code-instruct-4k
  147. 44.5 arcee-ai/AFM-4.5B-Base
  148. 44.4 allenai/Olmo-3-7B-Instruct
  149. 44.4 BSC-LT/salamandra-7b
  150. 44.4 Qwen/Qwen2.5-1.5B-Instruct
  151. 44.2 ibm-granite/granite-20b-code-instruct
  152. 44.2 Mistral-7B-Instruct-v0.1
  153. 42.7 allenai/Olmo-3-7B-Instruct-SFT
  154. 42.6 Gemma 3 4B
  155. 42.2 allenai/Olmo-3-7B-Instruct-DPO
  156. 41.6 AI-Sweden-Models/roberta-large-1160k
  157. 41.4 intfloat/multilingual-e5-large
  158. 39.7 meta-llama/Llama-2-70b-chat-hf
  159. 39.0 openGPT-X/Teuken-7B-instruct-commercial-v0.4
  160. 37.5 common-pile/comma-v0.1-2t
  161. 37.5 Phi-4
  162. 36.9 ibm-granite/granite-4.0-h-tiny
  163. 36.5 jhu-clsp/mmBERT-base
  164. 36.2 allenai/Olmo-3-7B-Think
  165. 36.0 ibm-granite/granite-3.0-3b-a800m-instruct
  166. 35.6 OLMo 2 32B Instruct
  167. 35.3 speakleash/Bielik-11B-v2.6-Instruct
  168. 35.2 deepcogito/cogito-v1-preview-llama-8B
  169. 34.5 FacebookAI/xlm-roberta-large
  170. 33.7 speakleash/Bielik-1.5B-v3.0-Instruct
  171. 33.7 Gemma 3 1B PT
  172. 33.5 GLM-4.5-Air
  173. 33.0 Qwen3-1.7B
  174. 32.9 Falcon3-3B-Instruct
  175. 32.5 MaLA-LM/emma-500-llama2-7b
  176. 32.0 Mistral-7B-Instruct-v0.2
  177. 31.7 Starling-LM-7B-beta
  178. 31.4 LiquidAI/LFM2-2.6B
  179. 31.2 NbAiLab/nb-roberta-base-ncc-plus-scandi-2e4
  180. 30.4 arcee-ai/Trinity-Nano-Preview
  181. 30.1 NbAiLab/nb-roberta-base-scandi-1e4
  182. 30.1 Gemma 2B
  183. 30.0 NbAiLab/borealis-4b-instruct-preview
  184. 29.7 ibm-granite/granite-3.0-3b-a800m-base
  185. 29.5 NbAiLab/nb-roberta-base-scandi
  186. 28.9 NbAiLab/nb-bert-large
  187. 27.7 vesteinn/FoBERT
  188. 27.5 SmolLM2 1.7B Instruct
  189. 27.1 EuroBERT/EuroBERT-610m
  190. 26.1 ltg/norbert3-base
  191. 26.0 Gemma 3 1B
  192. 25.9 jhu-clsp/mmBERT-small
  193. 25.8 Phi-4-Reasoning-plus
  194. 25.7 SmolLM2-1.7B
  195. 25.6 kyutai/helium-1-2b
  196. 24.4 NbAiLab/nb-roberta-base-ncc-plus-scandi-1e4
  197. 24.3 occiglot/occiglot-7b-eu5-instruct
  198. 23.8 meta-llama/Llama-4-Scout-17B-16E
  199. 23.8 meta-llama/Llama-2-13b-chat-hf
  200. 23.6 common-pile/comma-v0.1-1t
  201. 23.3 tencent/Hunyuan-7B-Instruct
  202. 23.2 DeepSeek-R1-Distill-Qwen-32B
  203. 23.1 Qwen/Qwen2.5-Math-7B
  204. 23.1 infly/OpenCoder-8B-Base
  205. 22.7 Qwen1.5-14B
  206. 22.2 DGurgurov/xlm-r_dan-latn
  207. 22.1 mideind/IceBERT-large
  208. 21.6 Llama 3.2 1B
  209. 21.3 tiiuae/falcon-11B
  210. 21.2 answerdotai/ModernBERT-large
  211. 21.0 KBLab/megatron-bert-large-swedish-cased-165k
  212. 20.4 tiiuae/Falcon3-3B-Base
  213. 20.3 Qwen3-0.6B
  214. 18.6 AI-Sweden-Models/bert-large-nordic-pile-1M-steps
  215. 18.4 arcee-ai/Trinity-Mini-Base
  216. 18.4 vesteinn/ScandiBERT-no-faroese
  217. 18.0 meta-llama/Llama-2-7b-chat-hf
  218. 17.9 arcee-ai/Trinity-Nano-Base
  219. 17.9 Nemotron 3 Nano 4B
  220. 17.9 yulan-team/YuLan-Mini-Instruct
  221. 17.9 answerdotai/ModernBERT-base
  222. 17.8 KBLab/megatron-bert-base-swedish-cased-600k
  223. 17.4 google/gemma-7b-it
  224. 17.2 AFM-4.5B
  225. 17.1 timpal0l/Mistral-7B-v0.1-flashback-v2
  226. 16.8 NbAiLab/notram-bert-norwegian-cased-080321
  227. 16.4 ltg/norbert3-small
  228. 16.3 PleIAs/Pleias-1.2b-Preview
  229. 15.2 BSC-LT/salamandra-7b-instruct
  230. 15.0 KBLab/megatron-bert-base-swedish-cased-125k
  231. 15.0 facebook/MobileLLM-R1-950M-base
  232. 14.8 google/gemma-2b-it
  233. 14.7 intfloat/multilingual-e5-base
  234. 14.6 EuroBERT/EuroBERT-210m
  235. 14.2 infly/OpenCoder-1.5B-Instruct
  236. 14.1 DeepSeek-R1-Distill-Qwen-14B
  237. 13.8 deepcogito/cogito-v1-preview-llama-3B
  238. 13.8 PleIAs/Pleias-3b-Preview
  239. 12.8 ibm-granite/granite-3b-code-base-2k
  240. 12.4 DiscoResearch/DiscoLM-70b
  241. 11.6 ltg/norbert2
  242. 10.8 utter-project/EuroMoE-2.6B-A0.6B-Preview
  243. 10.7 TurkuNLP/finnish-modernbert-large
  244. 10.7 DeepSeek-R1-Distill-Qwen-7B
  245. 10.6 TurkuNLP/finnish-modernbert-large-seq-len-1024-117300-annealed
  246. 10.4 deepseek-ai/DeepSeek-R1-Distill-Llama-8B
  247. 10.3 GLM-4.1V-9B-Thinking
  248. 10.0 ibm-granite/granite-3b-code-instruct-2k
  249. 9.97 tencent/Hunyuan-1.8B-Instruct
  250. 9.89 mideind/IceBERT-xlmr-ic3
  251. 9.78 baidu/ERNIE-4.5-0.3B-PT
  252. 9.67 DDSC/roberta-base-scandinavian
  253. 9.51 utter-project/EuroVLM-1.7B-Preview
  254. 9.43 AI-Sweden-Models/gpt-sw3-356m
  255. 8.38 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview
  256. 7.68 Phi-4-mini-reasoning
  257. 7.44 meta-llama/Llama-3.2-1B-Instruct
  258. 7.00 Qwen/Qwen1.5-0.5B-Chat
  259. 6.96 AI-Sweden-Models/gpt-sw3-356m-instruct
  260. 6.41 ibm-granite/granite-3.0-1b-a400m-instruct
  261. 6.23 TurkuNLP/finnish-modernbert-base
  262. 5.76 TurkuNLP/finnish-modernbert-tiny
  263. 5.73 KBLab/bert-base-swedish-cased-new
  264. 5.69 Qwen/Qwen2.5-Math-1.5B
  265. 5.46 NbAiLab/nb-roberta-base-ncc
  266. 5.46 LFM2-1.2B
  267. 5.24 EuropeanParliament/EUBERT
  268. 5.17 LiquidAI/LFM2-700M
  269. 4.95 microsoft/mdeberta-v3-base
  270. 4.67 DDSC/roberta-base-danish
  271. 4.57 mideind/IceBERT-igc
  272. 4.30 ibm-granite/granite-3.0-1b-a400m-base
  273. 4.27 Gemma 3 270M
  274. 4.12 SmolLM2-360M
  275. 4.11 HuggingFaceTB/SmolLM2-360M-Instruct
  276. 3.94 vesteinn/XLMR-ENIS
  277. 3.87 LiquidAI/LFM2.5-1.2B-Base
  278. 3.67 HPLT/hplt2c_isl_checkpoints
  279. 3.59 ltg/norbert3-xs
  280. 3.48 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base
  281. 3.41 HPLT/hplt2c_nno_checkpoints
  282. 3.33 Qwen/Qwen1.5-0.5B
  283. 3.20 HuggingFaceTB/SmolLM2-135M-Instruct
  284. 3.07 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct
  285. 2.88 Gemma 3 270M Base
  286. 2.87 HPLT/hplt2c_dan_checkpoints
  287. 2.85 LFM2.5-1.2B-Instruct
  288. 2.75 ltg/norbert
  289. 2.66 HPLT/hplt2c_swe_checkpoints
  290. 2.60 DeepSeek-R1-Distill-Qwen-1.5B
  291. 2.52 SmolLM2-135M
  292. 2.51 AI-Sweden-Models/gpt-sw3-126m-instruct
  293. 2.50 facebook/MobileLLM-R1-360M-base
  294. 2.50 PleIAs/Pleias-350m-Preview
  295. 2.45 HPLT/hplt2c_nob_checkpoints
  296. 2.38 Shekswess/trlm-135m
  297. 2.27 PleIAs/Monad
  298. 2.23 LiquidAI/LFM2-350M
  299. 1.85 infly/OpenCoder-1.5B-Base
  300. 1.83 PleIAs/Baguettotron@refs/pr/6
  301. 1.75 facebook/MobileLLM-R1-140M-base
  302. 1.69 AI-Sweden-Models/gpt-sw3-126m
  303. 1.61 Phi-2
  304. 1.20 facebook/MobileLLM-R1-360M
  305. 1.13 LFM2.5-1.2B-Thinking
  306. 1.05 mideind/IceBERT-ic3
  307. 1.03 mideind/IceBERT
  308. 0.99 PleIAs/Pleias-Pico
  309. 0.40 openai-community/gpt2-medium
  310. 0.29 Geotrend/distilbert-base-25lang-cased
  311. 0.28 distilbert/distilbert-base-multilingual-cased
  312. 0.23 openai-community/gpt2-large
  313. 0.15 facebook/MobileLLM-R1-950M
  314. 0.15 marksverdhei/modern-norbert3-small
  315. 0.13 openai-community/gpt2
  316. 0.07 NbAiLab/nb-bert-ncc-male2female
  317. 0.05 KBLab/bert-base-swedish-cased-alpha
  318. 0.05 Maltehb/danish-bert-botxo
  319. 0.05 marksverdhei/modern-norbert3-base
  320. 0.03 KBLab/sentence-bert-swedish-cased
  321. 0.01 facebook/MobileLLM-R1-140M
  322. 0.0 danish-foundation-models/encoder-small-v1
  323. 0.0 fresh-electra-small
  324. 0.0 fresh-xlm-roberta-base
  325. 0.0 KBLab/bert-base-swedish-cased
  326. 0.0 KBLab/roberta-base-swedish-cased
  327. 0.0 Maltehb/aelaectra-danish-electra-small-cased
  328. 0.0 Maltehb/aelaectra-danish-electra-small-uncased
  329. 0.0 marksverdhei/modern-norbert3-large
  330. 0.0 marksverdhei/modern-norbert3-xs
  331. 0.0 mideind/IceBERT-mC4-is
  332. 0.0 NbAiLab/nb-bert-base
  333. 0.0 NbAiLab/roberta_jan_128_ncc
  334. 0.0 neurocode/IsRoBERTa
  335. 0.0 vesteinn/DanskBERT
  336. 0.0 vesteinn/ScandiBERT
  1. 1 Mistral Small 3.2 79.9%
  2. 2 meta-llama/Llama-3.1-405B-Instruct-FP8 79.7%
  3. 3 mistralai/Magistral-Small-2509 79.6%
  4. 4 Mistral Small 3.1 79.5%
  5. 5 Llama 3.1-70B 79.3%
  6. 6 google/gemma-2-9b-it 78.6%
  7. 7 LumiOpen/Llama-Poro-2-70B-SFT 77.8%
  8. 8 mistralai/Mistral-Small-3.1-24B-Base-2503 77.5%
  9. 9 Cogito v1 Preview 70B 77.4%
  10. 10 Gemma 3 12B PT 77.3%
  11. 11 Gemma 3 27B PT 77.2%
  12. 12 deepcogito/cogito-v2-preview-llama-70B 76.8%
  13. 13 bullerwins/Magistral-Small-2506-fp8 76.7%
  14. 14 Apertus 70B 76.7%
  15. 15 deepcogito/cogito-v1-preview-qwen-32B 76.5%
  16. 16 speakleash/Bielik-11B-v3-Base-20250730 76.4%
  17. 17 deepcogito/cogito-v1-preview-qwen-14B 76.2%
  18. 18 TildeAI/TildeOpen-30b slow-tokenizer 75.9%
  19. 19 Qwen3-14B no-thinking 75.8%
  20. 20 Qwen/Qwen3-235B-A22B-Instruct-2507-FP8 75.7%
  21. 21 Voxtral Small 75.4%
  22. 22 Gemma 2 27B 75.3%
  23. 23 Gemma 3n E4B 74.2%
  24. 24 NovaSky-AI/Sky-T1-32B-Preview 74.2%
  25. 25 Qwen3-32B no-thinking 73.9%
  26. 26 mistralai/Ministral-3-8B-Reasoning-2512 73.6%
  27. 27 Gemma 2 9B 73.6%
  28. 28 Qwen3-Next-80B-A3B-Instruct 73.6%
  29. 29 Apertus 8B Instruct 73.4%
  30. 30 Llama 4 Scout 73.3%
  31. 31 Qwen/Qwen3-235B-A22B-Thinking-2507-FP8 73.1%
  32. 32 LumiOpen/Llama-Poro-2-70B-Instruct 73%
  33. 33 Qwen3-235B-A22B 72.9%
  34. 34 Llama 3.1-405B 72.7%
  35. 35 Gemma 3 27B 72.7%
  36. 36 swiss-ai/Apertus-8B-2509 72.5%
  37. 37 allenai/OLMo-2-0325-32B 72.5%
  38. 38 allenai/Olmo-3-1125-32B 72.5%
  39. 39 norallm/normistral-11b-long 72.3%
  40. 40 Gemma 7B 72.2%
  41. 41 utter-project/EuroLLM-22B-2512 72.1%
  42. 42 Llama 3.1-8B 71.8%
  43. 43 allenai/Llama-3.1-Tulu-3-8B-DPO 71.7%
  44. 44 Granite 4.0 H Small 71.5%
  45. 45 google/gemma-2-27b-it 71.1%
  46. 46 Reka Flash 3 71%
  47. 47 Aya Expanse 32B 70.9%
  48. 48 mistralai/Mistral-Small-24B-Instruct-2501 70.7%
  49. 49 mistralai/Mistral-Nemo-Instruct-2407 70.3%
  50. 50 Llama-3.1-Tulu-3-8B 70.3%
  51. 51 Gemma 3 12B 70.3%
  52. 52 Gemma 3 4B PT 70.3%
  53. 53 ThatsGroes/gemma-2-27b-it-FP8-Dynamic 70.2%
  54. 54 allenai/Llama-3.1-Tulu-3-8B-SFT 70.1%
  55. 55 Llama 3-8B 70%
  56. 56 meta-llama/Llama-3.1-8B-Instruct 69.7%
  57. 57 Gemma 3n E2B 69%
  58. 58 allenai/Llama-3.1-Tulu-3.1-8B 69%
  59. 59 norallm/normistral-11b-warm 68.8%
  60. 60 Qwen3-8B no-thinking 68.7%
  61. 61 mistralai/Ministral-3-3B-Base-2512 68.6%
  62. 62 ollama_chat/gemma3:27b-it-fp16 68.5%
  63. 63 BSC-LT/ALIA-40b 68.1%
  64. 64 QwQ-32B 67.8%
  65. 65 ai21labs/Jamba-v0.1 67.7%
  66. 66 syvai/llama3-da-base 67.5%
  67. 67 utter-project/EuroLLM-22B-Preview 67.1%
  68. 68 Llama 2-70B 67.1%
  69. 69 Yi-1.5-34B 66.9%
  70. 70 Qwen3-30B-A3B no-thinking 66.7%
  71. 71 Falcon3-10B-Instruct 66.6%
  72. 72 ollama_chat/gemma3:4b-it-q4_K_M 66.3%
  73. 73 qwen3-4b-instruct-2507 65.9%
  74. 74 NousResearch/DeepHermes-3-Llama-3-8B-Preview 65.7%
  75. 75 Qwen3-Next-80B-A3B 65.5%
  76. 76 upstage/SOLAR-10.7B-v1.0 65.1%
  77. 77 baidu/ERNIE-4.5-21B-A3B-PT 65%
  78. 78 CohereForAI/aya-23-35B 64.9%
  79. 79 Qwen3-4B-Thinking-2507 64.8%
  80. 80 Command R7B 64.2%
  81. 81 utter-project/EuroLLM-9B 64%
  82. 82 ollama_chat/hf.co/google/gemma-3-4b-it-qat-q4_0-gguf 63.6%
  83. 83 tiiuae/Falcon3-10B-Base 63.2%
  84. 84 mistralai/Voxtral-Mini-3B-2507 63.1%
  85. 85 01-ai/Yi-1.5-9B 62.6%
  86. 86 TrustLLMeu/baseline-7-8b_2-3t-tokens_llama 62.6%
  87. 87 neuralmagic/Sparse-Llama-3.1-8B-2of4 62.5%
  88. 88 allenai/OLMo-2-1124-13B-Instruct 62.3%
  89. 89 mii-llm/nesso-4B 62.2%
  90. 90 utter-project/EuroLLM-9B-Instruct 62.1%
  91. 91 allenai/OLMo-2-1124-13B 62%
  92. 92 Reka Flash 3.1 61.9%
  93. 93 gpt-oss-20b medium 61.6%
  94. 94 Phi-4-reasoning 61.5%
  95. 95 ibm-granite/granite-34b-code-base 61.4%
  96. 96 TrustLLMeu/baseline-7-8b_1t-tokens_llama 61.3%
  97. 97 Qwen3-4B no-thinking 61.2%
  98. 98 Mistral 7B v0.3 61.1%
  99. 99 Falcon3-7B-Instruct 61%
  100. 100 Nemotron 3 Nano 30B-A3B no-thinking 61%
  101. 101 LumiOpen/Viking-33B 60.4%
  102. 102 senseable/WestLake-7B-v2 60.3%
  103. 103 Qwen1.5-14B-Chat 60.1%
  104. 104 bineric/lynx-instruct-30b thinking 59.5%
  105. 105 Olmo 3.1 32B Instruct 58.4%
  106. 106 utter-project/EuroLLM-22B-Instruct-2512 58.3%
  107. 107 nvidia-nemotron-nano-9b-v2 58.1%
  108. 108 Granite-3.0-8B-Base 57.4%
  109. 109 Mistral 7B v0.1 57.1%
  110. 110 meta-llama/Llama-3.2-3B-Instruct 56.3%
  111. 111 utter-project/EuroLLM-22B-Instruct-Preview 56.2%
  112. 112 allenai/OLMo-2-1124-7B 55.9%
  113. 113 ibm-granite/granite-34b-code-instruct 55.7%
  114. 114 allenai/Olmo-3.1-32B-Instruct-SFT 55.6%
  115. 115 meta-llama/Llama-2-13b-hf 55.5%
  116. 116 allenai/OLMo-2-1124-7B-Instruct 55.2%
  117. 117 Granite-3.0-8B-Instruct 55.2%
  118. 118 allenai/Olmo-3.1-32B-Instruct-DPO 55%
  119. 119 norallm/normistral-11b-thinking 54.5%
  120. 120 tiiuae/Falcon3-7B-Base 54.4%
  121. 121 ibm-granite/granite-4.0-h-micro 54.2%
  122. 122 ibm-granite/granite-20b-code-base 54.2%
  123. 123 google/gemma-2-2b 53.7%
  124. 124 TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49 53.1%
  125. 125 allenai/Olmo-3-1025-7B 53.1%
  126. 126 deepseek-r1-0528-qwen3-8b 52.9%
  127. 127 speakleash/Bielik-11B-v2.3-Instruct 52.8%
  128. 128 speakleash/Bielik-4.5B-v3.0-Instruct 52.5%
  129. 129 LumiOpen/Viking-13B 52.1%
  130. 130 DeepSeek-R1-Distill-Llama-70B 51.2%
  131. 131 occiglot/occiglot-7b-eu5 51.2%
  132. 132 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6 51.2%
  133. 133 granite-4.0-micro 50.3%
  134. 134 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f zero-shot 49.9%
  135. 135 Gemma 2 2B IT 49.6%
  136. 136 HuggingFaceTB/SmolLM3-3B-Base 49.6%
  137. 137 ibm-granite/granite-8b-code-base-4k 49.4%
  138. 138 152334H/miqu-1-70b-sf 49.1%
  139. 139 TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894 48.4%
  140. 140 ltg/norbert3-large 48.2%
  141. 141 AI-Sweden-Models/ModernBERT-large 47%
  142. 142 infly/OpenCoder-8B-Instruct 46.8%
  143. 143 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660 zero-shot 45.9%
  144. 144 Ministral 3 3B Reasoning 45.6%
  145. 145 HuggingFaceTB/SmolLM3-3B 44.9%
  146. 146 ibm-granite/granite-8b-code-instruct-4k 44.9%
  147. 147 arcee-ai/AFM-4.5B-Base 44.5%
  148. 148 allenai/Olmo-3-7B-Instruct 44.4%
  149. 149 BSC-LT/salamandra-7b 44.4%
  150. 150 Qwen/Qwen2.5-1.5B-Instruct 44.4%
  151. 151 ibm-granite/granite-20b-code-instruct 44.2%
  152. 152 Mistral-7B-Instruct-v0.1 44.2%
  153. 153 allenai/Olmo-3-7B-Instruct-SFT 42.7%
  154. 154 Gemma 3 4B 42.6%
  155. 155 allenai/Olmo-3-7B-Instruct-DPO 42.2%
  156. 156 AI-Sweden-Models/roberta-large-1160k 41.6%
  157. 157 intfloat/multilingual-e5-large 41.4%
  158. 158 meta-llama/Llama-2-70b-chat-hf 39.7%
  159. 159 openGPT-X/Teuken-7B-instruct-commercial-v0.4 39%
  160. 160 common-pile/comma-v0.1-2t 37.5%
  161. 161 Phi-4 37.5%
  162. 162 ibm-granite/granite-4.0-h-tiny 36.9%
  163. 163 jhu-clsp/mmBERT-base 36.5%
  164. 164 allenai/Olmo-3-7B-Think 36.2%
  165. 165 ibm-granite/granite-3.0-3b-a800m-instruct 36%
  166. 166 OLMo 2 32B Instruct 35.6%
  167. 167 speakleash/Bielik-11B-v2.6-Instruct 35.3%
  168. 168 deepcogito/cogito-v1-preview-llama-8B 35.2%
  169. 169 FacebookAI/xlm-roberta-large 34.5%
  170. 170 speakleash/Bielik-1.5B-v3.0-Instruct 33.7%
  171. 171 Gemma 3 1B PT 33.7%
  172. 172 GLM-4.5-Air no-thinking 33.5%
  173. 173 Qwen3-1.7B 33%
  174. 174 Falcon3-3B-Instruct 32.9%
  175. 175 MaLA-LM/emma-500-llama2-7b 32.5%
  176. 176 Mistral-7B-Instruct-v0.2 32%
  177. 177 Starling-LM-7B-beta 31.7%
  178. 178 LiquidAI/LFM2-2.6B 31.4%
  179. 179 NbAiLab/nb-roberta-base-ncc-plus-scandi-2e4 31.2%
  180. 180 arcee-ai/Trinity-Nano-Preview 30.4%
  181. 181 NbAiLab/nb-roberta-base-scandi-1e4 30.1%
  182. 182 Gemma 2B 30.1%
  183. 183 NbAiLab/borealis-4b-instruct-preview 30%
  184. 184 ibm-granite/granite-3.0-3b-a800m-base 29.7%
  185. 185 NbAiLab/nb-roberta-base-scandi 29.5%
  186. 186 NbAiLab/nb-bert-large 28.9%
  187. 187 vesteinn/FoBERT 27.7%
  188. 188 SmolLM2 1.7B Instruct 27.5%
  189. 189 EuroBERT/EuroBERT-610m 27.1%
  190. 190 ltg/norbert3-base 26.1%
  191. 191 Gemma 3 1B 26%
  192. 192 jhu-clsp/mmBERT-small 25.9%
  193. 193 Phi-4-Reasoning-plus 25.8%
  194. 194 SmolLM2-1.7B 25.7%
  195. 195 kyutai/helium-1-2b 25.6%
  196. 196 NbAiLab/nb-roberta-base-ncc-plus-scandi-1e4 24.4%
  197. 197 occiglot/occiglot-7b-eu5-instruct 24.3%
  198. 198 meta-llama/Llama-4-Scout-17B-16E 23.8%
  199. 199 meta-llama/Llama-2-13b-chat-hf 23.8%
  200. 200 common-pile/comma-v0.1-1t 23.6%
  201. 201 tencent/Hunyuan-7B-Instruct 23.3%
  202. 202 DeepSeek-R1-Distill-Qwen-32B 23.2%
  203. 203 Qwen/Qwen2.5-Math-7B 23.1%
  204. 204 infly/OpenCoder-8B-Base 23.1%
  205. 205 Qwen1.5-14B 22.7%
  206. 206 DGurgurov/xlm-r_dan-latn 22.2%
  207. 207 mideind/IceBERT-large 22.1%
  208. 208 Llama 3.2 1B 21.6%
  209. 209 tiiuae/falcon-11B 21.3%
  210. 210 answerdotai/ModernBERT-large 21.2%
  211. 211 KBLab/megatron-bert-large-swedish-cased-165k 21%
  212. 212 tiiuae/Falcon3-3B-Base 20.4%
  213. 213 Qwen3-0.6B 20.3%
  214. 214 AI-Sweden-Models/bert-large-nordic-pile-1M-steps 18.6%
  215. 215 arcee-ai/Trinity-Mini-Base 18.4%
  216. 216 vesteinn/ScandiBERT-no-faroese 18.4%
  217. 217 meta-llama/Llama-2-7b-chat-hf 18%
  218. 218 arcee-ai/Trinity-Nano-Base 17.9%
  219. 219 Nemotron 3 Nano 4B 17.9%
  220. 220 yulan-team/YuLan-Mini-Instruct 17.9%
  221. 221 answerdotai/ModernBERT-base 17.9%
  222. 222 KBLab/megatron-bert-base-swedish-cased-600k 17.8%
  223. 223 google/gemma-7b-it 17.4%
  224. 224 AFM-4.5B 17.2%
  225. 225 timpal0l/Mistral-7B-v0.1-flashback-v2 17.1%
  226. 226 NbAiLab/notram-bert-norwegian-cased-080321 16.8%
  227. 227 ltg/norbert3-small 16.4%
  228. 228 PleIAs/Pleias-1.2b-Preview 16.3%
  229. 229 BSC-LT/salamandra-7b-instruct 15.2%
  230. 230 KBLab/megatron-bert-base-swedish-cased-125k 15%
  231. 231 facebook/MobileLLM-R1-950M-base 15%
  232. 232 google/gemma-2b-it 14.8%
  233. 233 intfloat/multilingual-e5-base 14.7%
  234. 234 EuroBERT/EuroBERT-210m 14.6%
  235. 235 infly/OpenCoder-1.5B-Instruct 14.2%
  236. 236 DeepSeek-R1-Distill-Qwen-14B 14.1%
  237. 237 deepcogito/cogito-v1-preview-llama-3B 13.8%
  238. 238 PleIAs/Pleias-3b-Preview 13.8%
  239. 239 ibm-granite/granite-3b-code-base-2k 12.8%
  240. 240 DiscoResearch/DiscoLM-70b 12.4%
  241. 241 ltg/norbert2 11.6%
  242. 242 utter-project/EuroMoE-2.6B-A0.6B-Preview 10.8%
  243. 243 TurkuNLP/finnish-modernbert-large 10.7%
  244. 244 DeepSeek-R1-Distill-Qwen-7B 10.7%
  245. 245 TurkuNLP/finnish-modernbert-large-seq-len-1024-117300-annealed 10.6%
  246. 246 deepseek-ai/DeepSeek-R1-Distill-Llama-8B 10.4%
  247. 247 GLM-4.1V-9B-Thinking 10.3%
  248. 248 ibm-granite/granite-3b-code-instruct-2k 10%
  249. 249 tencent/Hunyuan-1.8B-Instruct 9.97%
  250. 250 mideind/IceBERT-xlmr-ic3 9.89%
  251. 251 baidu/ERNIE-4.5-0.3B-PT 9.78%
  252. 252 DDSC/roberta-base-scandinavian 9.67%
  253. 253 utter-project/EuroVLM-1.7B-Preview 9.51%
  254. 254 AI-Sweden-Models/gpt-sw3-356m 9.43%
  255. 255 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview 8.38%
  256. 256 Phi-4-mini-reasoning 7.68%
  257. 257 meta-llama/Llama-3.2-1B-Instruct 7.44%
  258. 258 Qwen/Qwen1.5-0.5B-Chat 7.00%
  259. 259 AI-Sweden-Models/gpt-sw3-356m-instruct 6.96%
  260. 260 ibm-granite/granite-3.0-1b-a400m-instruct 6.41%
  261. 261 TurkuNLP/finnish-modernbert-base 6.23%
  262. 262 TurkuNLP/finnish-modernbert-tiny 5.76%
  263. 263 KBLab/bert-base-swedish-cased-new 5.73%
  264. 264 Qwen/Qwen2.5-Math-1.5B 5.69%
  265. 265 NbAiLab/nb-roberta-base-ncc 5.46%
  266. 266 LFM2-1.2B 5.46%
  267. 267 EuropeanParliament/EUBERT 5.24%
  268. 268 LiquidAI/LFM2-700M 5.17%
  269. 269 microsoft/mdeberta-v3-base 4.95%
  270. 270 DDSC/roberta-base-danish 4.67%
  271. 271 mideind/IceBERT-igc 4.57%
  272. 272 ibm-granite/granite-3.0-1b-a400m-base 4.30%
  273. 273 Gemma 3 270M 4.27%
  274. 274 SmolLM2-360M 4.12%
  275. 275 HuggingFaceTB/SmolLM2-360M-Instruct 4.11%
  276. 276 vesteinn/XLMR-ENIS 3.94%
  277. 277 LiquidAI/LFM2.5-1.2B-Base 3.87%
  278. 278 HPLT/hplt2c_isl_checkpoints 3.67%
  279. 279 ltg/norbert3-xs 3.59%
  280. 280 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base 3.48%
  281. 281 HPLT/hplt2c_nno_checkpoints 3.41%
  282. 282 Qwen/Qwen1.5-0.5B 3.33%
  283. 283 HuggingFaceTB/SmolLM2-135M-Instruct 3.20%
  284. 284 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct 3.07%
  285. 285 Gemma 3 270M Base 2.88%
  286. 286 HPLT/hplt2c_dan_checkpoints 2.87%
  287. 287 LFM2.5-1.2B-Instruct 2.85%
  288. 288 ltg/norbert 2.75%
  289. 289 HPLT/hplt2c_swe_checkpoints 2.66%
  290. 290 DeepSeek-R1-Distill-Qwen-1.5B 2.60%
  291. 291 SmolLM2-135M 2.52%
  292. 292 AI-Sweden-Models/gpt-sw3-126m-instruct 2.51%
  293. 293 facebook/MobileLLM-R1-360M-base 2.50%
  294. 294 PleIAs/Pleias-350m-Preview 2.50%
  295. 295 HPLT/hplt2c_nob_checkpoints 2.45%
  296. 296 Shekswess/trlm-135m 2.38%
  297. 297 PleIAs/Monad 2.27%
  298. 298 LiquidAI/LFM2-350M 2.23%
  299. 299 infly/OpenCoder-1.5B-Base 1.85%
  300. 300 PleIAs/Baguettotron@refs/pr/6 no-thinking 1.83%
  301. 301 facebook/MobileLLM-R1-140M-base 1.75%
  302. 302 AI-Sweden-Models/gpt-sw3-126m 1.69%
  303. 303 Phi-2 1.61%
  304. 304 facebook/MobileLLM-R1-360M 1.20%
  305. 305 LFM2.5-1.2B-Thinking 1.13%
  306. 306 mideind/IceBERT-ic3 1.05%
  307. 307 mideind/IceBERT 1.03%
  308. 308 PleIAs/Pleias-Pico 0.99%
  309. 309 openai-community/gpt2-medium 0.40%
  310. 310 Geotrend/distilbert-base-25lang-cased 0.29%
  311. 311 distilbert/distilbert-base-multilingual-cased 0.28%
  312. 312 openai-community/gpt2-large 0.23%
  313. 313 facebook/MobileLLM-R1-950M 0.15%
  314. 314 marksverdhei/modern-norbert3-small 0.15%
  315. 315 openai-community/gpt2 0.13%
  316. 316 NbAiLab/nb-bert-ncc-male2female 0.07%
  317. 317 KBLab/bert-base-swedish-cased-alpha 0.05%
  318. 318 Maltehb/danish-bert-botxo 0.05%
  319. 319 marksverdhei/modern-norbert3-base 0.05%
  320. 320 KBLab/sentence-bert-swedish-cased 0.03%
  321. 321 facebook/MobileLLM-R1-140M 0.01%
  322. 322 danish-foundation-models/encoder-small-v1 0%
  323. 323 fresh-electra-small 0%
  324. 324 fresh-xlm-roberta-base 0%
  325. 325 KBLab/bert-base-swedish-cased 0%
  326. 326 KBLab/roberta-base-swedish-cased 0%
  327. 327 Maltehb/aelaectra-danish-electra-small-cased 0%
  328. 328 Maltehb/aelaectra-danish-electra-small-uncased 0%
  329. 329 marksverdhei/modern-norbert3-large 0%
  330. 330 marksverdhei/modern-norbert3-xs 0%
  331. 331 mideind/IceBERT-mC4-is 0%
  332. 332 NbAiLab/nb-bert-base 0%
  333. 333 NbAiLab/roberta_jan_128_ncc 0%
  334. 334 neurocode/IsRoBERTa 0%
  335. 335 vesteinn/DanskBERT 0%
  336. 336 vesteinn/ScandiBERT 0%

Source: EuroEval’s leaderboard

91 models. Scroll sideways for the rest.

  1. 79.9 Mistral Small 3.2
  2. 79.5 Mistral Small 3.1
  3. 79.3 Llama 3.1-70B
  4. 77.4 Cogito v1 Preview 70B
  5. 77.3 Gemma 3 12B PT
  6. 77.2 Gemma 3 27B PT
  7. 76.7 Apertus 70B
  8. 75.8 Qwen3-14B
  9. 75.4 Voxtral Small
  10. 75.3 Gemma 2 27B
  11. 74.2 Gemma 3n E4B
  12. 73.9 Qwen3-32B
  13. 73.6 Gemma 2 9B
  14. 73.6 Qwen3-Next-80B-A3B-Instruct
  15. 73.4 Apertus 8B Instruct
  16. 73.3 Llama 4 Scout
  17. 72.9 Qwen3-235B-A22B
  18. 72.7 Llama 3.1-405B
  19. 72.7 Gemma 3 27B
  20. 72.2 Gemma 7B
  21. 71.8 Llama 3.1-8B
  22. 71.5 Granite 4.0 H Small
  23. 71.0 Reka Flash 3
  24. 70.9 Aya Expanse 32B
  25. 70.3 Llama-3.1-Tulu-3-8B
  26. 70.3 Gemma 3 12B
  27. 70.3 Gemma 3 4B PT
  28. 70.0 Llama 3-8B
  29. 69.0 Gemma 3n E2B
  30. 68.7 Qwen3-8B
  31. 67.8 QwQ-32B
  32. 67.1 Llama 2-70B
  33. 66.9 Yi-1.5-34B
  34. 66.7 Qwen3-30B-A3B
  35. 66.6 Falcon3-10B-Instruct
  36. 65.9 qwen3-4b-instruct-2507
  37. 65.5 Qwen3-Next-80B-A3B
  38. 64.8 Qwen3-4B-Thinking-2507
  39. 64.2 Command R7B
  40. 61.9 Reka Flash 3.1
  41. 61.6 gpt-oss-20b
  42. 61.5 Phi-4-reasoning
  43. 61.2 Qwen3-4B
  44. 61.1 Mistral 7B v0.3
  45. 61.0 Falcon3-7B-Instruct
  46. 61.0 Nemotron 3 Nano 30B-A3B
  47. 60.1 Qwen1.5-14B-Chat
  48. 58.4 Olmo 3.1 32B Instruct
  49. 58.1 nvidia-nemotron-nano-9b-v2
  50. 57.4 Granite-3.0-8B-Base
  51. 57.1 Mistral 7B v0.1
  52. 55.2 Granite-3.0-8B-Instruct
  53. 52.9 deepseek-r1-0528-qwen3-8b
  54. 51.2 DeepSeek-R1-Distill-Llama-70B
  55. 50.3 granite-4.0-micro
  56. 49.6 Gemma 2 2B IT
  57. 45.6 Ministral 3 3B Reasoning
  58. 44.2 Mistral-7B-Instruct-v0.1
  59. 42.6 Gemma 3 4B
  60. 37.5 Phi-4
  61. 35.6 OLMo 2 32B Instruct
  62. 33.7 Gemma 3 1B PT
  63. 33.5 GLM-4.5-Air
  64. 33.0 Qwen3-1.7B
  65. 32.9 Falcon3-3B-Instruct
  66. 32.0 Mistral-7B-Instruct-v0.2
  67. 31.7 Starling-LM-7B-beta
  68. 30.1 Gemma 2B
  69. 27.5 SmolLM2 1.7B Instruct
  70. 26.0 Gemma 3 1B
  71. 25.8 Phi-4-Reasoning-plus
  72. 25.7 SmolLM2-1.7B
  73. 23.2 DeepSeek-R1-Distill-Qwen-32B
  74. 22.7 Qwen1.5-14B
  75. 21.6 Llama 3.2 1B
  76. 20.3 Qwen3-0.6B
  77. 17.9 Nemotron 3 Nano 4B
  78. 17.2 AFM-4.5B
  79. 14.1 DeepSeek-R1-Distill-Qwen-14B
  80. 10.7 DeepSeek-R1-Distill-Qwen-7B
  81. 10.3 GLM-4.1V-9B-Thinking
  82. 7.68 Phi-4-mini-reasoning
  83. 5.46 LFM2-1.2B
  84. 4.27 Gemma 3 270M
  85. 4.12 SmolLM2-360M
  86. 2.88 Gemma 3 270M Base
  87. 2.85 LFM2.5-1.2B-Instruct
  88. 2.60 DeepSeek-R1-Distill-Qwen-1.5B
  89. 2.52 SmolLM2-135M
  90. 1.61 Phi-2
  91. 1.13 LFM2.5-1.2B-Thinking
  1. 1 Mistral Small 3.2 79.9%
  2. 2 Mistral Small 3.1 79.5%
  3. 3 Llama 3.1-70B 79.3%
  4. 4 Cogito v1 Preview 70B 77.4%
  5. 5 Gemma 3 12B PT 77.3%
  6. 6 Gemma 3 27B PT 77.2%
  7. 7 Apertus 70B 76.7%
  8. 8 Qwen3-14B no-thinking 75.8%
  9. 9 Voxtral Small 75.4%
  10. 10 Gemma 2 27B 75.3%
  11. 11 Gemma 3n E4B 74.2%
  12. 12 Qwen3-32B no-thinking 73.9%
  13. 13 Gemma 2 9B 73.6%
  14. 14 Qwen3-Next-80B-A3B-Instruct 73.6%
  15. 15 Apertus 8B Instruct 73.4%
  16. 16 Llama 4 Scout 73.3%
  17. 17 Qwen3-235B-A22B 72.9%
  18. 18 Llama 3.1-405B 72.7%
  19. 19 Gemma 3 27B 72.7%
  20. 20 Gemma 7B 72.2%
  21. 21 Llama 3.1-8B 71.8%
  22. 22 Granite 4.0 H Small 71.5%
  23. 23 Reka Flash 3 71%
  24. 24 Aya Expanse 32B 70.9%
  25. 25 Llama-3.1-Tulu-3-8B 70.3%
  26. 26 Gemma 3 12B 70.3%
  27. 27 Gemma 3 4B PT 70.3%
  28. 28 Llama 3-8B 70%
  29. 29 Gemma 3n E2B 69%
  30. 30 Qwen3-8B no-thinking 68.7%
  31. 31 QwQ-32B 67.8%
  32. 32 Llama 2-70B 67.1%
  33. 33 Yi-1.5-34B 66.9%
  34. 34 Qwen3-30B-A3B no-thinking 66.7%
  35. 35 Falcon3-10B-Instruct 66.6%
  36. 36 qwen3-4b-instruct-2507 65.9%
  37. 37 Qwen3-Next-80B-A3B 65.5%
  38. 38 Qwen3-4B-Thinking-2507 64.8%
  39. 39 Command R7B 64.2%
  40. 40 Reka Flash 3.1 61.9%
  41. 41 gpt-oss-20b medium 61.6%
  42. 42 Phi-4-reasoning 61.5%
  43. 43 Qwen3-4B no-thinking 61.2%
  44. 44 Mistral 7B v0.3 61.1%
  45. 45 Falcon3-7B-Instruct 61%
  46. 46 Nemotron 3 Nano 30B-A3B no-thinking 61%
  47. 47 Qwen1.5-14B-Chat 60.1%
  48. 48 Olmo 3.1 32B Instruct 58.4%
  49. 49 nvidia-nemotron-nano-9b-v2 58.1%
  50. 50 Granite-3.0-8B-Base 57.4%
  51. 51 Mistral 7B v0.1 57.1%
  52. 52 Granite-3.0-8B-Instruct 55.2%
  53. 53 deepseek-r1-0528-qwen3-8b 52.9%
  54. 54 DeepSeek-R1-Distill-Llama-70B 51.2%
  55. 55 granite-4.0-micro 50.3%
  56. 56 Gemma 2 2B IT 49.6%
  57. 57 Ministral 3 3B Reasoning 45.6%
  58. 58 Mistral-7B-Instruct-v0.1 44.2%
  59. 59 Gemma 3 4B 42.6%
  60. 60 Phi-4 37.5%
  61. 61 OLMo 2 32B Instruct 35.6%
  62. 62 Gemma 3 1B PT 33.7%
  63. 63 GLM-4.5-Air no-thinking 33.5%
  64. 64 Qwen3-1.7B 33%
  65. 65 Falcon3-3B-Instruct 32.9%
  66. 66 Mistral-7B-Instruct-v0.2 32%
  67. 67 Starling-LM-7B-beta 31.7%
  68. 68 Gemma 2B 30.1%
  69. 69 SmolLM2 1.7B Instruct 27.5%
  70. 70 Gemma 3 1B 26%
  71. 71 Phi-4-Reasoning-plus 25.8%
  72. 72 SmolLM2-1.7B 25.7%
  73. 73 DeepSeek-R1-Distill-Qwen-32B 23.2%
  74. 74 Qwen1.5-14B 22.7%
  75. 75 Llama 3.2 1B 21.6%
  76. 76 Qwen3-0.6B 20.3%
  77. 77 Nemotron 3 Nano 4B 17.9%
  78. 78 AFM-4.5B 17.2%
  79. 79 DeepSeek-R1-Distill-Qwen-14B 14.1%
  80. 80 DeepSeek-R1-Distill-Qwen-7B 10.7%
  81. 81 GLM-4.1V-9B-Thinking 10.3%
  82. 82 Phi-4-mini-reasoning 7.68%
  83. 83 LFM2-1.2B 5.46%
  84. 84 Gemma 3 270M 4.27%
  85. 85 SmolLM2-360M 4.12%
  86. 86 Gemma 3 270M Base 2.88%
  87. 87 LFM2.5-1.2B-Instruct 2.85%
  88. 88 DeepSeek-R1-Distill-Qwen-1.5B 2.60%
  89. 89 SmolLM2-135M 2.52%
  90. 90 Phi-2 1.61%
  91. 91 LFM2.5-1.2B-Thinking 1.13%

Source: EuroEval’s leaderboard

#EntryScore ±SettingDateLinkSource
1 Mistral Small 3.2 79.9% ±0.7 Jun 2025 hf.co Official leaderboard
2 meta-llama/Llama-3.1-405B-Instruct-FP8 open weights 79.7% ±0.5 hf.co Official leaderboard
3 mistralai/Magistral-Small-2509 open weights 79.6% ±0.5 hf.co Official leaderboard
4 Mistral Small 3.1 79.5% ±0.9 Mar 2025 hf.co Official leaderboard
5 Llama 3.1-70B 79.3% ±0.6 Jul 2024 hf.co Official leaderboard
6 google/gemma-2-9b-it open weights 78.6% ±0.4 hf.co Official leaderboard
7 LumiOpen/Llama-Poro-2-70B-SFT open weights 77.8% ±0.5 hf.co Official leaderboard
8 mistralai/Mistral-Small-3.1-24B-Base-2503 open weights 77.5% ±1.1 hf.co Official leaderboard
9 Cogito v1 Preview 70B 77.4% ±0.5 Apr 2025 hf.co Official leaderboard
10 Gemma 3 12B PT 77.3% ±0.8 Mar 2025 hf.co Official leaderboard
11 Gemma 3 27B PT 77.2% ±1 Mar 2025 hf.co Official leaderboard
12 deepcogito/cogito-v2-preview-llama-70B open weights 76.8% ±0.4 hf.co Official leaderboard
13 bullerwins/Magistral-Small-2506-fp8 open weights 76.7% ±0.8 hf.co Official leaderboard
14 Apertus 70B 76.7% ±0.9 Sep 2025 hf.co Official leaderboard
15 deepcogito/cogito-v1-preview-qwen-32B open weights 76.5% ±0.8 hf.co Official leaderboard
16 speakleash/Bielik-11B-v3-Base-20250730 open weights 76.4% ±1 hf.co Official leaderboard
17 deepcogito/cogito-v1-preview-qwen-14B open weights 76.2% ±0.7 hf.co Official leaderboard
18 TildeAI/TildeOpen-30b open weights 75.9% ±0.7 slow-tokenizer hf.co Official leaderboard
19 Qwen3-14B 75.8% ±1.3 no-thinking Apr 2025 hf.co Official leaderboard
20 Qwen/Qwen3-235B-A22B-Instruct-2507-FP8 closed weights 75.7% ±0.7 hf.co Official leaderboard
21 Voxtral Small 75.4% ±2.7 Jul 2025 hf.co Official leaderboard
22 Gemma 2 27B 75.3% ±0.9 Jun 2024 hf.co Official leaderboard
23 Gemma 3n E4B 74.2% ±0.8 Jun 2025 hf.co Official leaderboard
24 NovaSky-AI/Sky-T1-32B-Preview open weights 74.2% ±0.7 hf.co Official leaderboard
25 Qwen3-32B 73.9% ±1.9 no-thinking Apr 2025 hf.co Official leaderboard
26 mistralai/Ministral-3-8B-Reasoning-2512 open weights 73.6% ±1.1 hf.co Official leaderboard
27 Gemma 2 9B 73.6% ±1.2 Jun 2024 hf.co Official leaderboard
28 Qwen3-Next-80B-A3B-Instruct 73.6% ±0.4 Sep 2025 hf.co Official leaderboard
29 Apertus 8B Instruct 73.4% ±0.8 Sep 2025 hf.co Official leaderboard
30 Llama 4 Scout 73.3% ±0.7 Apr 2025 hf.co Official leaderboard
31 Qwen/Qwen3-235B-A22B-Thinking-2507-FP8 closed weights 73.1% ±0.4 hf.co Official leaderboard
32 LumiOpen/Llama-Poro-2-70B-Instruct open weights 73% ±0.5 hf.co Official leaderboard
33 Qwen3-235B-A22B 72.9% ±0.4 Apr 2025 hf.co Official leaderboard
34 Llama 3.1-405B 72.7% ±3.8 Jul 2024 hf.co Official leaderboard
35 Gemma 3 27B 72.7% ±0.7 Mar 2025 hf.co Official leaderboard
36 swiss-ai/Apertus-8B-2509 open weights 72.5% ±0.6 hf.co Official leaderboard
37 allenai/OLMo-2-0325-32B open weights 72.5% ±0.5 hf.co Official leaderboard
38 allenai/Olmo-3-1125-32B open weights 72.5% ±0.4 hf.co Official leaderboard
39 norallm/normistral-11b-long open weights 72.3% ±1.2 hf.co Official leaderboard
40 Gemma 7B 72.2% ±0.7 Feb 2024 hf.co Official leaderboard
41 utter-project/EuroLLM-22B-2512 open weights 72.1% ±0.9 hf.co Official leaderboard
42 Llama 3.1-8B 71.8% ±0.7 Jul 2024 hf.co Official leaderboard
43 allenai/Llama-3.1-Tulu-3-8B-DPO open weights 71.7% ±0.7 hf.co Official leaderboard
44 Granite 4.0 H Small 71.5% ±0.9 Oct 2025 hf.co Official leaderboard
45 google/gemma-2-27b-it open weights 71.1% ±0.7 hf.co Official leaderboard
46 Reka Flash 3 71% ±0.5 Mar 2025 hf.co Official leaderboard
47 Aya Expanse 32B 70.9% ±1.5 Oct 2024 hf.co Official leaderboard
48 mistralai/Mistral-Small-24B-Instruct-2501 open weights 70.7% ±2.5 hf.co Official leaderboard
49 mistralai/Mistral-Nemo-Instruct-2407 open weights 70.3% ±0.7 hf.co Official leaderboard
50 Llama-3.1-Tulu-3-8B 70.3% ±0.8 Nov 2024 hf.co Official leaderboard
51 Gemma 3 12B 70.3% ±0.6 Mar 2025 hf.co Official leaderboard
52 Gemma 3 4B PT 70.3% ±0.9 Mar 2025 hf.co Official leaderboard
53 ThatsGroes/gemma-2-27b-it-FP8-Dynamic open weights 70.2% ±0.8 hf.co Official leaderboard
54 allenai/Llama-3.1-Tulu-3-8B-SFT open weights 70.1% ±0.9 hf.co Official leaderboard
55 Llama 3-8B 70% ±1 Apr 2024 hf.co Official leaderboard
56 meta-llama/Llama-3.1-8B-Instruct open weights 69.7% ±0.8 hf.co Official leaderboard
57 Gemma 3n E2B 69% ±0.6 Jun 2025 hf.co Official leaderboard
58 allenai/Llama-3.1-Tulu-3.1-8B open weights 69% ±0.6 hf.co Official leaderboard
59 norallm/normistral-11b-warm open weights 68.8% ±3.3 hf.co Official leaderboard
60 Qwen3-8B 68.7% ±1 no-thinking Apr 2025 hf.co Official leaderboard
61 mistralai/Ministral-3-3B-Base-2512 open weights 68.6% ±0.9 hf.co Official leaderboard
62 ollama_chat/gemma3:27b-it-fp16 open weights 68.5% ±0.5 ollama.com Official leaderboard
63 BSC-LT/ALIA-40b open weights 68.1% ±0.6 hf.co Official leaderboard
64 QwQ-32B 67.8% ±0.6 Mar 2025 hf.co Official leaderboard
65 ai21labs/Jamba-v0.1 open weights 67.7% ±0.7 hf.co Official leaderboard
66 syvai/llama3-da-base closed weights 67.5% ±1 Official leaderboard
67 utter-project/EuroLLM-22B-Preview open weights 67.1% ±0.8 hf.co Official leaderboard
68 Llama 2-70B 67.1% ±1.1 Jul 2023 hf.co Official leaderboard
69 Yi-1.5-34B 66.9% ±0.8 May 2024 hf.co Official leaderboard
70 Qwen3-30B-A3B 66.7% ±1.3 no-thinking Apr 2025 hf.co Official leaderboard
71 Falcon3-10B-Instruct 66.6% ±1 Dec 2024 hf.co Official leaderboard
72 ollama_chat/gemma3:4b-it-q4_K_M open weights 66.3% ±0.5 ollama.com Official leaderboard
73 qwen3-4b-instruct-2507 65.9% ±0.7 Aug 2025 hf.co Official leaderboard
74 NousResearch/DeepHermes-3-Llama-3-8B-Preview open weights 65.7% ±0.9 hf.co Official leaderboard
75 Qwen3-Next-80B-A3B 65.5% ±0.4 Sep 2025 hf.co Official leaderboard
76 upstage/SOLAR-10.7B-v1.0 open weights 65.1% ±1 hf.co Official leaderboard
77 baidu/ERNIE-4.5-21B-A3B-PT open weights 65% ±4.3 hf.co Official leaderboard
78 CohereForAI/aya-23-35B open weights 64.9% ±0.9 hf.co Official leaderboard
79 Qwen3-4B-Thinking-2507 64.8% ±0.5 Aug 2025 hf.co Official leaderboard
80 Command R7B 64.2% ±1.3 Dec 2024 hf.co Official leaderboard
81 utter-project/EuroLLM-9B open weights 64% ±0.8 hf.co Official leaderboard
82 ollama_chat/hf.co/google/gemma-3-4b-it-qat-q4_0-gguf open weights 63.6% ±0.6 ollama.com Official leaderboard
83 tiiuae/Falcon3-10B-Base open weights 63.2% ±0.5 hf.co Official leaderboard
84 mistralai/Voxtral-Mini-3B-2507 open weights 63.1% ±0.7 hf.co Official leaderboard
85 01-ai/Yi-1.5-9B open weights 62.6% ±0.9 hf.co Official leaderboard
86 TrustLLMeu/baseline-7-8b_2-3t-tokens_llama open weights 62.6% ±1 hf.co Official leaderboard
87 neuralmagic/Sparse-Llama-3.1-8B-2of4 open weights 62.5% ±0.8 hf.co Official leaderboard
88 allenai/OLMo-2-1124-13B-Instruct open weights 62.3% ±0.9 hf.co Official leaderboard
89 mii-llm/nesso-4B open weights 62.2% ±1.3 hf.co Official leaderboard
90 utter-project/EuroLLM-9B-Instruct open weights 62.1% ±1.1 hf.co Official leaderboard
91 allenai/OLMo-2-1124-13B open weights 62% ±0.7 hf.co Official leaderboard
92 Reka Flash 3.1 61.9% ±0.6 May 2025 hf.co Official leaderboard
93 gpt-oss-20b 61.6% ±0.5 medium Aug 2025 hf.co Official leaderboard
94 Phi-4-reasoning 61.5% ±6.8 Apr 2025 hf.co Official leaderboard
95 ibm-granite/granite-34b-code-base open weights 61.4% ±0.8 hf.co Official leaderboard
96 TrustLLMeu/baseline-7-8b_1t-tokens_llama open weights 61.3% ±0.8 hf.co Official leaderboard
97 Qwen3-4B 61.2% ±0.8 no-thinking Apr 2025 hf.co Official leaderboard
98 Mistral 7B v0.3 61.1% ±0.5 Sep 2023 hf.co Official leaderboard
99 Falcon3-7B-Instruct 61% ±0.8 Dec 2024 hf.co Official leaderboard
100 Nemotron 3 Nano 30B-A3B 61% ±1.5 no-thinking Dec 2025 hf.co Official leaderboard
101 LumiOpen/Viking-33B open weights 60.4% ±0.9 hf.co Official leaderboard
102 senseable/WestLake-7B-v2 open weights 60.3% ±0.7 hf.co Official leaderboard
103 Qwen1.5-14B-Chat 60.1% ±2.5 Feb 2024 hf.co Official leaderboard
104 bineric/lynx-instruct-30b open weights 59.5% ±1.8 thinking hf.co Official leaderboard
105 Olmo 3.1 32B Instruct 58.4% ±4.5 Dec 2025 hf.co Official leaderboard
106 utter-project/EuroLLM-22B-Instruct-2512 open weights 58.3% ±2.6 hf.co Official leaderboard
107 nvidia-nemotron-nano-9b-v2 58.1% ±0.7 Aug 2025 hf.co Official leaderboard
108 Granite-3.0-8B-Base 57.4% ±0.9 Oct 2024 hf.co Official leaderboard
109 Mistral 7B v0.1 57.1% ±0.9 Sep 2023 hf.co Official leaderboard
110 meta-llama/Llama-3.2-3B-Instruct open weights 56.3% ±0.6 hf.co Official leaderboard
111 utter-project/EuroLLM-22B-Instruct-Preview open weights 56.2% ±2.9 hf.co Official leaderboard
112 allenai/OLMo-2-1124-7B open weights 55.9% ±0.7 hf.co Official leaderboard
113 ibm-granite/granite-34b-code-instruct open weights 55.7% ±1.2 hf.co Official leaderboard
114 allenai/Olmo-3.1-32B-Instruct-SFT open weights 55.6% ±5.8 hf.co Official leaderboard
115 meta-llama/Llama-2-13b-hf open weights 55.5% ±1.2 hf.co Official leaderboard
116 allenai/OLMo-2-1124-7B-Instruct open weights 55.2% ±1 hf.co Official leaderboard
117 Granite-3.0-8B-Instruct 55.2% ±0.8 Oct 2024 hf.co Official leaderboard
118 allenai/Olmo-3.1-32B-Instruct-DPO open weights 55% ±5.8 hf.co Official leaderboard
119 norallm/normistral-11b-thinking open weights 54.5% ±5.3 hf.co Official leaderboard
120 tiiuae/Falcon3-7B-Base open weights 54.4% ±0.9 hf.co Official leaderboard
121 ibm-granite/granite-4.0-h-micro open weights 54.2% ±3.1 hf.co Official leaderboard
122 ibm-granite/granite-20b-code-base open weights 54.2% ±1.3 hf.co Official leaderboard
123 google/gemma-2-2b open weights 53.7% ±1.2 hf.co Official leaderboard
124 TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49 open weights 53.1% ±0.7 hf.co Official leaderboard
125 allenai/Olmo-3-1025-7B open weights 53.1% ±1.1 hf.co Official leaderboard
126 deepseek-r1-0528-qwen3-8b 52.9% ±0.7 May 2025 hf.co Official leaderboard
127 speakleash/Bielik-11B-v2.3-Instruct open weights 52.8% ±4 hf.co Official leaderboard
128 speakleash/Bielik-4.5B-v3.0-Instruct open weights 52.5% ±3.5 hf.co Official leaderboard
129 LumiOpen/Viking-13B open weights 52.1% ±1.5 hf.co Official leaderboard
130 DeepSeek-R1-Distill-Llama-70B 51.2% ±0.8 Jan 2025 hf.co Official leaderboard
131 occiglot/occiglot-7b-eu5 open weights 51.2% ±1.1 hf.co Official leaderboard
132 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6 open weights 51.2% ±0.5 hf.co Official leaderboard
133 granite-4.0-micro 50.3% ±2.5 Oct 2025 hf.co Official leaderboard
134 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f open weights 49.9% ±0.4 zero-shot hf.co Official leaderboard
135 Gemma 2 2B IT 49.6% ±1.6 Jul 2024 hf.co Official leaderboard
136 HuggingFaceTB/SmolLM3-3B-Base open weights 49.6% ±0.9 hf.co Official leaderboard
137 ibm-granite/granite-8b-code-base-4k open weights 49.4% ±0.8 hf.co Official leaderboard
138 152334H/miqu-1-70b-sf open weights 49.1% ±1.6 hf.co Official leaderboard
139 TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894 open weights 48.4% ±1.3 hf.co Official leaderboard
140 ltg/norbert3-large open weights 48.2% ±1 hf.co Official leaderboard
141 AI-Sweden-Models/ModernBERT-large open weights 47% ±0.7 hf.co Official leaderboard
142 infly/OpenCoder-8B-Instruct open weights 46.8% ±1.1 hf.co Official leaderboard
143 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660 open weights 45.9% ±0.2 zero-shot hf.co Official leaderboard
144 Ministral 3 3B Reasoning 45.6% ±3.5 Dec 2025 hf.co Official leaderboard
145 HuggingFaceTB/SmolLM3-3B open weights 44.9% ±2.6 hf.co Official leaderboard
146 ibm-granite/granite-8b-code-instruct-4k open weights 44.9% ±0.7 hf.co Official leaderboard
147 arcee-ai/AFM-4.5B-Base open weights 44.5% ±0.9 hf.co Official leaderboard
148 allenai/Olmo-3-7B-Instruct open weights 44.4% ±0.6 hf.co Official leaderboard
149 BSC-LT/salamandra-7b open weights 44.4% ±1.4 hf.co Official leaderboard
150 Qwen/Qwen2.5-1.5B-Instruct open weights 44.4% ±0.7 hf.co Official leaderboard
151 ibm-granite/granite-20b-code-instruct open weights 44.2% ±1.1 hf.co Official leaderboard
152 Mistral-7B-Instruct-v0.1 44.2% ±1.2 Sep 2023 hf.co Official leaderboard
153 allenai/Olmo-3-7B-Instruct-SFT open weights 42.7% ±1.1 hf.co Official leaderboard
154 Gemma 3 4B 42.6% ±4.2 Mar 2025 hf.co Official leaderboard
155 allenai/Olmo-3-7B-Instruct-DPO open weights 42.2% ±1 hf.co Official leaderboard
156 AI-Sweden-Models/roberta-large-1160k open weights 41.6% ±1.9 hf.co Official leaderboard
157 intfloat/multilingual-e5-large open weights 41.4% ±1.5 hf.co Official leaderboard
158 meta-llama/Llama-2-70b-chat-hf open weights 39.7% ±1.9 hf.co Official leaderboard
159 openGPT-X/Teuken-7B-instruct-commercial-v0.4 open weights 39% ±2.1 hf.co Official leaderboard
160 common-pile/comma-v0.1-2t open weights 37.5% ±1.2 hf.co Official leaderboard
161 Phi-4 37.5% ±1 Dec 2024 hf.co Official leaderboard
162 ibm-granite/granite-4.0-h-tiny open weights 36.9% ±2.4 hf.co Official leaderboard
163 jhu-clsp/mmBERT-base open weights 36.5% ±1.1 hf.co Official leaderboard
164 allenai/Olmo-3-7B-Think open weights 36.2% ±1.8 hf.co Official leaderboard
165 ibm-granite/granite-3.0-3b-a800m-instruct open weights 36% ±0.8 hf.co Official leaderboard
166 OLMo 2 32B Instruct 35.6% ±0.7 Mar 2025 hf.co Official leaderboard
167 speakleash/Bielik-11B-v2.6-Instruct open weights 35.3% ±2.3 hf.co Official leaderboard
168 deepcogito/cogito-v1-preview-llama-8B open weights 35.2% ±1.4 hf.co Official leaderboard
169 FacebookAI/xlm-roberta-large open weights 34.5% ±0.9 hf.co Official leaderboard
170 speakleash/Bielik-1.5B-v3.0-Instruct open weights 33.7% ±1.6 hf.co Official leaderboard
171 Gemma 3 1B PT 33.7% ±1.4 Mar 2025 hf.co Official leaderboard
172 GLM-4.5-Air 33.5% ±0.9 no-thinking Jul 2025 hf.co Official leaderboard
173 Qwen3-1.7B 33% ±2.2 Apr 2025 hf.co Official leaderboard
174 Falcon3-3B-Instruct 32.9% ±1.1 Dec 2024 hf.co Official leaderboard
175 MaLA-LM/emma-500-llama2-7b open weights 32.5% ±3.6 hf.co Official leaderboard
176 Mistral-7B-Instruct-v0.2 32% ±1.4 Dec 2023 hf.co Official leaderboard
177 Starling-LM-7B-beta 31.7% ±1.9 Mar 2024 hf.co Official leaderboard
178 LiquidAI/LFM2-2.6B open weights 31.4% ±1 hf.co Official leaderboard
179 NbAiLab/nb-roberta-base-ncc-plus-scandi-2e4 open weights 31.2% ±0.9 hf.co Official leaderboard
180 arcee-ai/Trinity-Nano-Preview open weights 30.4% ±5.7 hf.co Official leaderboard
181 NbAiLab/nb-roberta-base-scandi-1e4 open weights 30.1% ±2.3 hf.co Official leaderboard
182 Gemma 2B 30.1% ±3.3 Feb 2024 hf.co Official leaderboard
183 NbAiLab/borealis-4b-instruct-preview open weights 30% ±1.1 hf.co Official leaderboard
184 ibm-granite/granite-3.0-3b-a800m-base open weights 29.7% ±0.7 hf.co Official leaderboard
185 NbAiLab/nb-roberta-base-scandi open weights 29.5% ±1 hf.co Official leaderboard
186 NbAiLab/nb-bert-large open weights 28.9% ±1.5 hf.co Official leaderboard
187 vesteinn/FoBERT open weights 27.7% ±1.5 hf.co Official leaderboard
188 SmolLM2 1.7B Instruct 27.5% ±0.8 Oct 2024 hf.co Official leaderboard
189 EuroBERT/EuroBERT-610m open weights 27.1% ±1 hf.co Official leaderboard
190 ltg/norbert3-base open weights 26.1% ±1.1 hf.co Official leaderboard
191 Gemma 3 1B 26% ±0.6 Mar 2025 ollama.com Official leaderboard
192 jhu-clsp/mmBERT-small open weights 25.9% ±1 hf.co Official leaderboard
193 Phi-4-Reasoning-plus 25.8% ±5.6 Apr 2025 hf.co Official leaderboard
194 SmolLM2-1.7B 25.7% ±1.3 Oct 2024 hf.co Official leaderboard
195 kyutai/helium-1-2b open weights 25.6% ±1.8 hf.co Official leaderboard
196 NbAiLab/nb-roberta-base-ncc-plus-scandi-1e4 open weights 24.4% ±4.2 hf.co Official leaderboard
197 occiglot/occiglot-7b-eu5-instruct open weights 24.3% ±0.5 hf.co Official leaderboard
198 meta-llama/Llama-4-Scout-17B-16E open weights 23.8% ±1.8 hf.co Official leaderboard
199 meta-llama/Llama-2-13b-chat-hf open weights 23.8% ±1.7 hf.co Official leaderboard
200 common-pile/comma-v0.1-1t open weights 23.6% ±1.7 hf.co Official leaderboard
201 tencent/Hunyuan-7B-Instruct open weights 23.3% ±0.5 hf.co Official leaderboard
202 DeepSeek-R1-Distill-Qwen-32B 23.2% ±1 Jan 2025 hf.co Official leaderboard
203 Qwen/Qwen2.5-Math-7B open weights 23.1% ±2.6 hf.co Official leaderboard
204 infly/OpenCoder-8B-Base open weights 23.1% ±0.4 hf.co Official leaderboard
205 Qwen1.5-14B 22.7% ±0.8 Feb 2024 hf.co Official leaderboard
206 DGurgurov/xlm-r_dan-latn open weights 22.2% ±1.3 hf.co Official leaderboard
207 mideind/IceBERT-large open weights 22.1% ±2.5 hf.co Official leaderboard
208 Llama 3.2 1B 21.6% ±1.8 Sep 2024 hf.co Official leaderboard
209 tiiuae/falcon-11B open weights 21.3% ±0.7 hf.co Official leaderboard
210 answerdotai/ModernBERT-large open weights 21.2% ±0.9 hf.co Official leaderboard
211 KBLab/megatron-bert-large-swedish-cased-165k open weights 21% ±0.9 hf.co Official leaderboard
212 tiiuae/Falcon3-3B-Base open weights 20.4% ±1.4 hf.co Official leaderboard
213 Qwen3-0.6B 20.3% ±0.8 Apr 2025 hf.co Official leaderboard
214 AI-Sweden-Models/bert-large-nordic-pile-1M-steps open weights 18.6% ±0.9 hf.co Official leaderboard
215 arcee-ai/Trinity-Mini-Base open weights 18.4% ±1.7 hf.co Official leaderboard
216 vesteinn/ScandiBERT-no-faroese open weights 18.4% ±2.4 hf.co Official leaderboard
217 meta-llama/Llama-2-7b-chat-hf open weights 18% ±1.6 hf.co Official leaderboard
218 arcee-ai/Trinity-Nano-Base open weights 17.9% ±0.7 hf.co Official leaderboard
219 Nemotron 3 Nano 4B 17.9% ±3.4 Mar 2026 hf.co Official leaderboard
220 yulan-team/YuLan-Mini-Instruct open weights 17.9% ±1.2 hf.co Official leaderboard
221 answerdotai/ModernBERT-base open weights 17.9% ±1 hf.co Official leaderboard
222 KBLab/megatron-bert-base-swedish-cased-600k open weights 17.8% ±1.1 hf.co Official leaderboard
223 google/gemma-7b-it open weights 17.4% ±0.9 hf.co Official leaderboard
224 AFM-4.5B 17.2% ±0.6 Jul 2025 hf.co Official leaderboard
225 timpal0l/Mistral-7B-v0.1-flashback-v2 open weights 17.1% ±0.7 hf.co Official leaderboard
226 NbAiLab/notram-bert-norwegian-cased-080321 open weights 16.8% ±1.2 hf.co Official leaderboard
227 ltg/norbert3-small open weights 16.4% ±2.2 hf.co Official leaderboard
228 PleIAs/Pleias-1.2b-Preview open weights 16.3% ±1.2 hf.co Official leaderboard
229 BSC-LT/salamandra-7b-instruct open weights 15.2% ±0.3 hf.co Official leaderboard
230 KBLab/megatron-bert-base-swedish-cased-125k open weights 15% ±0.9 hf.co Official leaderboard
231 facebook/MobileLLM-R1-950M-base open weights 15% ±1.2 hf.co Official leaderboard
232 google/gemma-2b-it open weights 14.8% ±0.9 hf.co Official leaderboard
233 intfloat/multilingual-e5-base open weights 14.7% ±2.6 hf.co Official leaderboard
234 EuroBERT/EuroBERT-210m open weights 14.6% ±3.1 hf.co Official leaderboard
235 infly/OpenCoder-1.5B-Instruct open weights 14.2% ±0.6 hf.co Official leaderboard
236 DeepSeek-R1-Distill-Qwen-14B 14.1% ±0.6 Jan 2025 hf.co Official leaderboard
237 deepcogito/cogito-v1-preview-llama-3B open weights 13.8% ±0.7 hf.co Official leaderboard
238 PleIAs/Pleias-3b-Preview open weights 13.8% ±0.8 hf.co Official leaderboard
239 ibm-granite/granite-3b-code-base-2k open weights 12.8% ±5.2 hf.co Official leaderboard
240 DiscoResearch/DiscoLM-70b open weights 12.4% ±3.3 hf.co Official leaderboard
241 ltg/norbert2 open weights 11.6% ±0.5 hf.co Official leaderboard
242 utter-project/EuroMoE-2.6B-A0.6B-Preview open weights 10.8% ±0.8 hf.co Official leaderboard
243 TurkuNLP/finnish-modernbert-large open weights 10.7% ±1.5 hf.co Official leaderboard
244 DeepSeek-R1-Distill-Qwen-7B 10.7% ±0.4 Jan 2025 hf.co Official leaderboard
245 TurkuNLP/finnish-modernbert-large-seq-len-1024-117300-annealed open weights 10.6% ±1.4 hf.co Official leaderboard
246 deepseek-ai/DeepSeek-R1-Distill-Llama-8B open weights 10.4% ±0.2 hf.co Official leaderboard
247 GLM-4.1V-9B-Thinking 10.3% ±0.4 Jul 2025 hf.co Official leaderboard
248 ibm-granite/granite-3b-code-instruct-2k open weights 10% ±4.1 hf.co Official leaderboard
249 tencent/Hunyuan-1.8B-Instruct open weights 9.97% ±0.4 hf.co Official leaderboard
250 mideind/IceBERT-xlmr-ic3 open weights 9.89% ±2.3 hf.co Official leaderboard
251 baidu/ERNIE-4.5-0.3B-PT open weights 9.78% ±0.8 hf.co Official leaderboard
252 DDSC/roberta-base-scandinavian open weights 9.67% ±1.7 hf.co Official leaderboard
253 utter-project/EuroVLM-1.7B-Preview open weights 9.51% ±0.4 hf.co Official leaderboard
254 AI-Sweden-Models/gpt-sw3-356m open weights 9.43% ±3 hf.co Official leaderboard
255 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview open weights 8.38% ±0.5 hf.co Official leaderboard
256 Phi-4-mini-reasoning 7.68% ±0.3 Apr 2025 hf.co Official leaderboard
257 meta-llama/Llama-3.2-1B-Instruct open weights 7.44% ±0.6 hf.co Official leaderboard
258 Qwen/Qwen1.5-0.5B-Chat open weights 7.00% ±0.5 hf.co Official leaderboard
259 AI-Sweden-Models/gpt-sw3-356m-instruct open weights 6.96% ±2.3 hf.co Official leaderboard
260 ibm-granite/granite-3.0-1b-a400m-instruct open weights 6.41% ±0.8 hf.co Official leaderboard
261 TurkuNLP/finnish-modernbert-base open weights 6.23% ±0.4 hf.co Official leaderboard
262 TurkuNLP/finnish-modernbert-tiny open weights 5.76% ±0.3 hf.co Official leaderboard
263 KBLab/bert-base-swedish-cased-new open weights 5.73% ±1.5 hf.co Official leaderboard
264 Qwen/Qwen2.5-Math-1.5B open weights 5.69% ±0.5 hf.co Official leaderboard
265 NbAiLab/nb-roberta-base-ncc open weights 5.46% ±2.5 hf.co Official leaderboard
266 LFM2-1.2B 5.46% ±0.3 Jul 2025 hf.co Official leaderboard
267 EuropeanParliament/EUBERT open weights 5.24% ±1 hf.co Official leaderboard
268 LiquidAI/LFM2-700M open weights 5.17% ±0.3 hf.co Official leaderboard
269 microsoft/mdeberta-v3-base open weights 4.95% ±2.3 hf.co Official leaderboard
270 DDSC/roberta-base-danish open weights 4.67% ±1.3 hf.co Official leaderboard
271 mideind/IceBERT-igc open weights 4.57% ±1.6 hf.co Official leaderboard
272 ibm-granite/granite-3.0-1b-a400m-base open weights 4.30% ±0.4 hf.co Official leaderboard
273 Gemma 3 270M 4.27% ±0.2 Aug 2025 hf.co Official leaderboard
274 SmolLM2-360M 4.12% ±0.4 Oct 2024 hf.co Official leaderboard
275 HuggingFaceTB/SmolLM2-360M-Instruct open weights 4.11% ±0.2 hf.co Official leaderboard
276 vesteinn/XLMR-ENIS open weights 3.94% ±1.2 hf.co Official leaderboard
277 LiquidAI/LFM2.5-1.2B-Base open weights 3.87% ±0.2 hf.co Official leaderboard
278 HPLT/hplt2c_isl_checkpoints open weights 3.67% ±0.2 hf.co Official leaderboard
279 ltg/norbert3-xs open weights 3.59% ±1.5 hf.co Official leaderboard
280 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base open weights 3.48% ±0.5 hf.co Official leaderboard
281 HPLT/hplt2c_nno_checkpoints open weights 3.41% ±0.2 hf.co Official leaderboard
282 Qwen/Qwen1.5-0.5B open weights 3.33% ±0.3 hf.co Official leaderboard
283 HuggingFaceTB/SmolLM2-135M-Instruct open weights 3.20% ±0.1 hf.co Official leaderboard
284 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct open weights 3.07% ±0.5 hf.co Official leaderboard
285 Gemma 3 270M Base 2.88% ±0.4 Aug 2025 hf.co Official leaderboard
286 HPLT/hplt2c_dan_checkpoints open weights 2.87% ±0.2 hf.co Official leaderboard
287 LFM2.5-1.2B-Instruct 2.85% ±0.1 Jan 2026 hf.co Official leaderboard
288 ltg/norbert open weights 2.75% ±0.2 hf.co Official leaderboard
289 HPLT/hplt2c_swe_checkpoints open weights 2.66% ±0.2 hf.co Official leaderboard
290 DeepSeek-R1-Distill-Qwen-1.5B 2.60% ±0.1 Jan 2025 hf.co Official leaderboard
291 SmolLM2-135M 2.52% ±0.4 Oct 2024 hf.co Official leaderboard
292 AI-Sweden-Models/gpt-sw3-126m-instruct open weights 2.51% ±0.9 hf.co Official leaderboard
293 facebook/MobileLLM-R1-360M-base open weights 2.50% ±0.4 hf.co Official leaderboard
294 PleIAs/Pleias-350m-Preview open weights 2.50% ±0.3 hf.co Official leaderboard
295 HPLT/hplt2c_nob_checkpoints open weights 2.45% ±0.2 hf.co Official leaderboard
296 Shekswess/trlm-135m open weights 2.38% ±0.3 hf.co Official leaderboard
297 PleIAs/Monad open weights 2.27% ±0.1 hf.co Official leaderboard
298 LiquidAI/LFM2-350M open weights 2.23% ±0.1 hf.co Official leaderboard
299 infly/OpenCoder-1.5B-Base open weights 1.85% ±0.6 hf.co Official leaderboard
300 PleIAs/Baguettotron@refs/pr/6 open weights 1.83% ±0.1 no-thinking hf.co Official leaderboard
301 facebook/MobileLLM-R1-140M-base open weights 1.75% ±0.3 hf.co Official leaderboard
302 AI-Sweden-Models/gpt-sw3-126m open weights 1.69% ±0.6 hf.co Official leaderboard
303 Phi-2 1.61% ±0.8 Dec 2023 hf.co Official leaderboard
304 facebook/MobileLLM-R1-360M open weights 1.20% ±0.1 hf.co Official leaderboard
305 LFM2.5-1.2B-Thinking 1.13% ±0.1 Jan 2026 hf.co Official leaderboard
306 mideind/IceBERT-ic3 open weights 1.05% ±0.6 hf.co Official leaderboard
307 mideind/IceBERT open weights 1.03% ±1 hf.co Official leaderboard
308 PleIAs/Pleias-Pico open weights 0.99% ±0.4 hf.co Official leaderboard
309 openai-community/gpt2-medium open weights 0.40% ±0.2 hf.co Official leaderboard
310 Geotrend/distilbert-base-25lang-cased open weights 0.29% ±0.2 hf.co Official leaderboard
311 distilbert/distilbert-base-multilingual-cased open weights 0.28% ±0.1 hf.co Official leaderboard
312 openai-community/gpt2-large open weights 0.23% ±0.1 hf.co Official leaderboard
313 facebook/MobileLLM-R1-950M open weights 0.15% ±3.9 hf.co Official leaderboard
314 marksverdhei/modern-norbert3-small open weights 0.15% ±6.1 hf.co Official leaderboard
315 openai-community/gpt2 open weights 0.13% ±7 hf.co Official leaderboard
316 NbAiLab/nb-bert-ncc-male2female open weights 0.07% ±5.5 hf.co Official leaderboard
317 KBLab/bert-base-swedish-cased-alpha open weights 0.05% ±4.2 hf.co Official leaderboard
318 Maltehb/danish-bert-botxo open weights 0.05% ±1.9 hf.co Official leaderboard
319 marksverdhei/modern-norbert3-base open weights 0.05% ±5.1 hf.co Official leaderboard
320 KBLab/sentence-bert-swedish-cased open weights 0.03% ±2.4 hf.co Official leaderboard
321 facebook/MobileLLM-R1-140M open weights 0.01% ±1 hf.co Official leaderboard
322 danish-foundation-models/encoder-small-v1 open weights 0% hf.co Official leaderboard
323 fresh-electra-small open weights 0% Official leaderboard
324 fresh-xlm-roberta-base open weights 0% Official leaderboard
325 KBLab/bert-base-swedish-cased open weights 0% hf.co Official leaderboard
326 KBLab/roberta-base-swedish-cased open weights 0% hf.co Official leaderboard
327 Maltehb/aelaectra-danish-electra-small-cased open weights 0% hf.co Official leaderboard
328 Maltehb/aelaectra-danish-electra-small-uncased open weights 0% hf.co Official leaderboard
329 marksverdhei/modern-norbert3-large open weights 0% hf.co Official leaderboard
330 marksverdhei/modern-norbert3-xs open weights 0% hf.co Official leaderboard
331 mideind/IceBERT-mC4-is open weights 0% hf.co Official leaderboard
332 NbAiLab/nb-bert-base open weights 0% hf.co Official leaderboard
333 NbAiLab/roberta_jan_128_ncc open weights 0% hf.co Official leaderboard
334 neurocode/IsRoBERTa open weights 0% ±0.3 hf.co Official leaderboard
335 vesteinn/DanskBERT open weights 0% hf.co Official leaderboard
336 vesteinn/ScandiBERT open weights 0% hf.co Official leaderboard
#ModelScore ±SettingDateLinkSource
1 Mistral Small 3.2 79.9% ±0.7 Jun 2025 hf.co Official leaderboard
2 Mistral Small 3.1 79.5% ±0.9 Mar 2025 hf.co Official leaderboard
3 Llama 3.1-70B 79.3% ±0.6 Jul 2024 hf.co Official leaderboard
4 Cogito v1 Preview 70B 77.4% ±0.5 Apr 2025 hf.co Official leaderboard
5 Gemma 3 12B PT 77.3% ±0.8 Mar 2025 hf.co Official leaderboard
6 Gemma 3 27B PT 77.2% ±1 Mar 2025 hf.co Official leaderboard
7 Apertus 70B 76.7% ±0.9 Sep 2025 hf.co Official leaderboard
8 Qwen3-14B 75.8% ±1.3 no-thinking Apr 2025 hf.co Official leaderboard
9 Voxtral Small 75.4% ±2.7 Jul 2025 hf.co Official leaderboard
10 Gemma 2 27B 75.3% ±0.9 Jun 2024 hf.co Official leaderboard
11 Gemma 3n E4B 74.2% ±0.8 Jun 2025 hf.co Official leaderboard
12 Qwen3-32B 73.9% ±1.9 no-thinking Apr 2025 hf.co Official leaderboard
13 Gemma 2 9B 73.6% ±1.2 Jun 2024 hf.co Official leaderboard
14 Qwen3-Next-80B-A3B-Instruct 73.6% ±0.4 Sep 2025 hf.co Official leaderboard
15 Apertus 8B Instruct 73.4% ±0.8 Sep 2025 hf.co Official leaderboard
16 Llama 4 Scout 73.3% ±0.7 Apr 2025 hf.co Official leaderboard
17 Qwen3-235B-A22B 72.9% ±0.4 Apr 2025 hf.co Official leaderboard
18 Llama 3.1-405B 72.7% ±3.8 Jul 2024 hf.co Official leaderboard
19 Gemma 3 27B 72.7% ±0.7 Mar 2025 hf.co Official leaderboard
20 Gemma 7B 72.2% ±0.7 Feb 2024 hf.co Official leaderboard
21 Llama 3.1-8B 71.8% ±0.7 Jul 2024 hf.co Official leaderboard
22 Granite 4.0 H Small 71.5% ±0.9 Oct 2025 hf.co Official leaderboard
23 Reka Flash 3 71% ±0.5 Mar 2025 hf.co Official leaderboard
24 Aya Expanse 32B 70.9% ±1.5 Oct 2024 hf.co Official leaderboard
25 Llama-3.1-Tulu-3-8B 70.3% ±0.8 Nov 2024 hf.co Official leaderboard
26 Gemma 3 12B 70.3% ±0.6 Mar 2025 hf.co Official leaderboard
27 Gemma 3 4B PT 70.3% ±0.9 Mar 2025 hf.co Official leaderboard
28 Llama 3-8B 70% ±1 Apr 2024 hf.co Official leaderboard
29 Gemma 3n E2B 69% ±0.6 Jun 2025 hf.co Official leaderboard
30 Qwen3-8B 68.7% ±1 no-thinking Apr 2025 hf.co Official leaderboard
31 QwQ-32B 67.8% ±0.6 Mar 2025 hf.co Official leaderboard
32 Llama 2-70B 67.1% ±1.1 Jul 2023 hf.co Official leaderboard
33 Yi-1.5-34B 66.9% ±0.8 May 2024 hf.co Official leaderboard
34 Qwen3-30B-A3B 66.7% ±1.3 no-thinking Apr 2025 hf.co Official leaderboard
35 Falcon3-10B-Instruct 66.6% ±1 Dec 2024 hf.co Official leaderboard
36 qwen3-4b-instruct-2507 65.9% ±0.7 Aug 2025 hf.co Official leaderboard
37 Qwen3-Next-80B-A3B 65.5% ±0.4 Sep 2025 hf.co Official leaderboard
38 Qwen3-4B-Thinking-2507 64.8% ±0.5 Aug 2025 hf.co Official leaderboard
39 Command R7B 64.2% ±1.3 Dec 2024 hf.co Official leaderboard
40 Reka Flash 3.1 61.9% ±0.6 May 2025 hf.co Official leaderboard
41 gpt-oss-20b 61.6% ±0.5 medium Aug 2025 hf.co Official leaderboard
42 Phi-4-reasoning 61.5% ±6.8 Apr 2025 hf.co Official leaderboard
43 Qwen3-4B 61.2% ±0.8 no-thinking Apr 2025 hf.co Official leaderboard
44 Mistral 7B v0.3 61.1% ±0.5 Sep 2023 hf.co Official leaderboard
45 Falcon3-7B-Instruct 61% ±0.8 Dec 2024 hf.co Official leaderboard
46 Nemotron 3 Nano 30B-A3B 61% ±1.5 no-thinking Dec 2025 hf.co Official leaderboard
47 Qwen1.5-14B-Chat 60.1% ±2.5 Feb 2024 hf.co Official leaderboard
48 Olmo 3.1 32B Instruct 58.4% ±4.5 Dec 2025 hf.co Official leaderboard
49 nvidia-nemotron-nano-9b-v2 58.1% ±0.7 Aug 2025 hf.co Official leaderboard
50 Granite-3.0-8B-Base 57.4% ±0.9 Oct 2024 hf.co Official leaderboard
51 Mistral 7B v0.1 57.1% ±0.9 Sep 2023 hf.co Official leaderboard
52 Granite-3.0-8B-Instruct 55.2% ±0.8 Oct 2024 hf.co Official leaderboard
53 deepseek-r1-0528-qwen3-8b 52.9% ±0.7 May 2025 hf.co Official leaderboard
54 DeepSeek-R1-Distill-Llama-70B 51.2% ±0.8 Jan 2025 hf.co Official leaderboard
55 granite-4.0-micro 50.3% ±2.5 Oct 2025 hf.co Official leaderboard
56 Gemma 2 2B IT 49.6% ±1.6 Jul 2024 hf.co Official leaderboard
57 Ministral 3 3B Reasoning 45.6% ±3.5 Dec 2025 hf.co Official leaderboard
58 Mistral-7B-Instruct-v0.1 44.2% ±1.2 Sep 2023 hf.co Official leaderboard
59 Gemma 3 4B 42.6% ±4.2 Mar 2025 hf.co Official leaderboard
60 Phi-4 37.5% ±1 Dec 2024 hf.co Official leaderboard
61 OLMo 2 32B Instruct 35.6% ±0.7 Mar 2025 hf.co Official leaderboard
62 Gemma 3 1B PT 33.7% ±1.4 Mar 2025 hf.co Official leaderboard
63 GLM-4.5-Air 33.5% ±0.9 no-thinking Jul 2025 hf.co Official leaderboard
64 Qwen3-1.7B 33% ±2.2 Apr 2025 hf.co Official leaderboard
65 Falcon3-3B-Instruct 32.9% ±1.1 Dec 2024 hf.co Official leaderboard
66 Mistral-7B-Instruct-v0.2 32% ±1.4 Dec 2023 hf.co Official leaderboard
67 Starling-LM-7B-beta 31.7% ±1.9 Mar 2024 hf.co Official leaderboard
68 Gemma 2B 30.1% ±3.3 Feb 2024 hf.co Official leaderboard
69 SmolLM2 1.7B Instruct 27.5% ±0.8 Oct 2024 hf.co Official leaderboard
70 Gemma 3 1B 26% ±0.6 Mar 2025 ollama.com Official leaderboard
71 Phi-4-Reasoning-plus 25.8% ±5.6 Apr 2025 hf.co Official leaderboard
72 SmolLM2-1.7B 25.7% ±1.3 Oct 2024 hf.co Official leaderboard
73 DeepSeek-R1-Distill-Qwen-32B 23.2% ±1 Jan 2025 hf.co Official leaderboard
74 Qwen1.5-14B 22.7% ±0.8 Feb 2024 hf.co Official leaderboard
75 Llama 3.2 1B 21.6% ±1.8 Sep 2024 hf.co Official leaderboard
76 Qwen3-0.6B 20.3% ±0.8 Apr 2025 hf.co Official leaderboard
77 Nemotron 3 Nano 4B 17.9% ±3.4 Mar 2026 hf.co Official leaderboard
78 AFM-4.5B 17.2% ±0.6 Jul 2025 hf.co Official leaderboard
79 DeepSeek-R1-Distill-Qwen-14B 14.1% ±0.6 Jan 2025 hf.co Official leaderboard
80 DeepSeek-R1-Distill-Qwen-7B 10.7% ±0.4 Jan 2025 hf.co Official leaderboard
81 GLM-4.1V-9B-Thinking 10.3% ±0.4 Jul 2025 hf.co Official leaderboard
82 Phi-4-mini-reasoning 7.68% ±0.3 Apr 2025 hf.co Official leaderboard
83 LFM2-1.2B 5.46% ±0.3 Jul 2025 hf.co Official leaderboard
84 Gemma 3 270M 4.27% ±0.2 Aug 2025 hf.co Official leaderboard
85 SmolLM2-360M 4.12% ±0.4 Oct 2024 hf.co Official leaderboard
86 Gemma 3 270M Base 2.88% ±0.4 Aug 2025 hf.co Official leaderboard
87 LFM2.5-1.2B-Instruct 2.85% ±0.1 Jan 2026 hf.co Official leaderboard
88 DeepSeek-R1-Distill-Qwen-1.5B 2.60% ±0.1 Jan 2025 hf.co Official leaderboard
89 SmolLM2-135M 2.52% ±0.4 Oct 2024 hf.co Official leaderboard
90 Phi-2 1.61% ±0.8 Dec 2023 hf.co Official leaderboard
91 LFM2.5-1.2B-Thinking 1.13% ±0.1 Jan 2026 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
75.8%
Qwen3-8B
no-thinking
68.7%
gpt-oss-20b
low
61.6%
Qwen3-4B
no-thinking
61.2%
nvidia-nemotron-nano-9b-v2
no-thinking
58.1%
GLM-4.5-Air
thinking no-thinking
33.5%
Qwen3-1.7B
no-thinking
33%
Qwen3-0.6B
no-thinking
20.3%

Source: EuroEval’s leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Mistral Small 3.2 79.9%
Mistral Small 3.1 79.5%
Llama 3.1-70B 79.3%
Cogito v1 Preview 70B 77.4%
Gemma 3 12B PT 77.3%
Gemma 3 27B PT 77.2%
Apertus 70B 76.7%
Qwen3-14B no-thinking 75.8%
Voxtral Small 75.4%
Gemma 2 27B 75.3%
Gemma 3n E4B 74.2%
Qwen3-32B no-thinking 73.9%
Gemma 2 9B 73.6%
Qwen3-Next-80B-A3B-Instruct 73.6%
Apertus 8B Instruct 73.4%
Llama 4 Scout 73.3%
Apertus 70B 73.1%
Qwen3-235B-A22B 72.9%
Llama 3.1-405B 72.7%
Gemma 3 27B 72.7%
Gemma 7B 72.2%
Qwen3-14B 72%
Llama 3.1-8B 71.8%
Granite 4.0 H Small 71.5%
Reka Flash 3 71%
Aya Expanse 32B 70.9%
Llama-3.1-Tulu-3-8B 70.3%
Gemma 3 12B 70.3%
Gemma 3 4B PT 70.3%
Llama 3-8B 70%
Gemma 3n E2B 69%
Qwen3-8B no-thinking 68.7%
Gemma 3n E4B 67.9%
QwQ-32B 67.8%
Llama 2-70B 67.1%
Yi-1.5-34B 66.9%
Qwen3-30B-A3B no-thinking 66.7%
Falcon3-10B-Instruct 66.6%
qwen3-4b-instruct-2507 65.9%
Qwen3-Next-80B-A3B 65.5%
Qwen3-4B-Thinking-2507 64.8%
Qwen3-8B 64.6%
Command R7B 64.2%
Reka Flash 3.1 61.9%
gpt-oss-20b medium 61.6%
Phi-4-reasoning 61.5%
Qwen3-4B no-thinking 61.2%
Qwen3-4B 61.2%
Mistral 7B v0.3 61.1%
Falcon3-7B-Instruct 61%
Nemotron 3 Nano 30B-A3B no-thinking 61%
Qwen1.5-14B-Chat 60.1%
Olmo 3.1 32B Instruct 58.4%
nvidia-nemotron-nano-9b-v2 58.1%
gpt-oss-20b low 57.6%
gpt-oss-20b high 57.6%
Granite-3.0-8B-Base 57.4%
Mistral 7B v0.1 57.1%
Granite-3.0-8B-Instruct 55.2%
deepseek-r1-0528-qwen3-8b 52.9%
DeepSeek-R1-Distill-Llama-70B 51.2%
granite-4.0-micro 50.3%
Gemma 2 2B IT 49.6%
Ministral 3 3B Reasoning 45.6%
Mistral-7B-Instruct-v0.1 44.2%
Gemma 3 4B 42.6%
Phi-4 37.5%
OLMo 2 32B Instruct 35.6%
Gemma 3 1B PT 33.7%
GLM-4.5-Air no-thinking 33.5%
Qwen3-1.7B 33%
Falcon3-3B-Instruct 32.9%
Mistral-7B-Instruct-v0.2 32%
Starling-LM-7B-beta 31.7%
Gemma 2B 30.1%
Qwen3-1.7B no-thinking 28.4%
SmolLM2 1.7B Instruct 27.5%
GLM-4.5-Air thinking 27.2%
Gemma 3 1B 26%
Phi-4-Reasoning-plus 25.8%
SmolLM2-1.7B 25.7%
DeepSeek-R1-Distill-Qwen-32B 23.2%
Qwen1.5-14B 22.7%
Llama 3.2 1B 21.6%
Qwen3-0.6B 20.3%
Nemotron 3 Nano 4B 17.9%
AFM-4.5B 17.2%
DeepSeek-R1-Distill-Qwen-14B 14.1%
Gemma 3 1B 10.7%
DeepSeek-R1-Distill-Qwen-7B 10.7%
GLM-4.1V-9B-Thinking 10.3%
Qwen3-0.6B no-thinking 9.38%
Phi-4-mini-reasoning 7.68%
LFM2-1.2B 5.46%
Gemma 3 270M 4.27%
SmolLM2-360M 4.12%
Gemma 3 270M Base 2.88%
LFM2.5-1.2B-Instruct 2.85%
DeepSeek-R1-Distill-Qwen-1.5B 2.60%
SmolLM2-135M 2.52%
Phi-2 1.61%
LFM2.5-1.2B-Thinking 1.13%
nvidia-nemotron-nano-9b-v2 no-thinking 0.63%

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
75.8%
Qwen3-8B
no-thinking
68.7%
gpt-oss-20b
low
61.6%
Qwen3-4B
no-thinking
61.2%
nvidia-nemotron-nano-9b-v2
no-thinking
58.1%
GLM-4.5-Air
thinking no-thinking
33.5%
Qwen3-1.7B
no-thinking
33%
Qwen3-0.6B
no-thinking
20.3%

Source: EuroEval’s leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Mistral Small 3.2 79.9%
Mistral Small 3.1 79.5%
Llama 3.1-70B 79.3%
Cogito v1 Preview 70B 77.4%
Gemma 3 12B PT 77.3%
Gemma 3 27B PT 77.2%
Apertus 70B 76.7%
Qwen3-14B no-thinking 75.8%
Voxtral Small 75.4%
Gemma 2 27B 75.3%
Gemma 3n E4B 74.2%
Qwen3-32B no-thinking 73.9%
Gemma 2 9B 73.6%
Qwen3-Next-80B-A3B-Instruct 73.6%
Apertus 8B Instruct 73.4%
Llama 4 Scout 73.3%
Apertus 70B 73.1%
Qwen3-235B-A22B 72.9%
Llama 3.1-405B 72.7%
Gemma 3 27B 72.7%
Gemma 7B 72.2%
Qwen3-14B 72%
Llama 3.1-8B 71.8%
Granite 4.0 H Small 71.5%
Reka Flash 3 71%
Aya Expanse 32B 70.9%
Llama-3.1-Tulu-3-8B 70.3%
Gemma 3 12B 70.3%
Gemma 3 4B PT 70.3%
Llama 3-8B 70%
Gemma 3n E2B 69%
Qwen3-8B no-thinking 68.7%
Gemma 3n E4B 67.9%
QwQ-32B 67.8%
Llama 2-70B 67.1%
Yi-1.5-34B 66.9%
Qwen3-30B-A3B no-thinking 66.7%
Falcon3-10B-Instruct 66.6%
qwen3-4b-instruct-2507 65.9%
Qwen3-Next-80B-A3B 65.5%
Qwen3-4B-Thinking-2507 64.8%
Qwen3-8B 64.6%
Command R7B 64.2%
Reka Flash 3.1 61.9%
gpt-oss-20b medium 61.6%
Phi-4-reasoning 61.5%
Qwen3-4B no-thinking 61.2%
Qwen3-4B 61.2%
Mistral 7B v0.3 61.1%
Falcon3-7B-Instruct 61%
Nemotron 3 Nano 30B-A3B no-thinking 61%
Qwen1.5-14B-Chat 60.1%
Olmo 3.1 32B Instruct 58.4%
nvidia-nemotron-nano-9b-v2 58.1%
gpt-oss-20b low 57.6%
gpt-oss-20b high 57.6%
Granite-3.0-8B-Base 57.4%
Mistral 7B v0.1 57.1%
Granite-3.0-8B-Instruct 55.2%
deepseek-r1-0528-qwen3-8b 52.9%
DeepSeek-R1-Distill-Llama-70B 51.2%
granite-4.0-micro 50.3%
Gemma 2 2B IT 49.6%
Ministral 3 3B Reasoning 45.6%
Mistral-7B-Instruct-v0.1 44.2%
Gemma 3 4B 42.6%
Phi-4 37.5%
OLMo 2 32B Instruct 35.6%
Gemma 3 1B PT 33.7%
GLM-4.5-Air no-thinking 33.5%
Qwen3-1.7B 33%
Falcon3-3B-Instruct 32.9%
Mistral-7B-Instruct-v0.2 32%
Starling-LM-7B-beta 31.7%
Gemma 2B 30.1%
Qwen3-1.7B no-thinking 28.4%
SmolLM2 1.7B Instruct 27.5%
GLM-4.5-Air thinking 27.2%
Gemma 3 1B 26%
Phi-4-Reasoning-plus 25.8%
SmolLM2-1.7B 25.7%
DeepSeek-R1-Distill-Qwen-32B 23.2%
Qwen1.5-14B 22.7%
Llama 3.2 1B 21.6%
Qwen3-0.6B 20.3%
Nemotron 3 Nano 4B 17.9%
AFM-4.5B 17.2%
DeepSeek-R1-Distill-Qwen-14B 14.1%
Gemma 3 1B 10.7%
DeepSeek-R1-Distill-Qwen-7B 10.7%
GLM-4.1V-9B-Thinking 10.3%
Qwen3-0.6B no-thinking 9.38%
Phi-4-mini-reasoning 7.68%
LFM2-1.2B 5.46%
Gemma 3 270M 4.27%
SmolLM2-360M 4.12%
Gemma 3 270M Base 2.88%
LFM2.5-1.2B-Instruct 2.85%
DeepSeek-R1-Distill-Qwen-1.5B 2.60%
SmolLM2-135M 2.52%
Phi-2 1.61%
LFM2.5-1.2B-Thinking 1.13%
nvidia-nemotron-nano-9b-v2 no-thinking 0.63%

3 models have held number one on EuroEval’s leaderboard since its release in February 2025

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

Meta Mistral AI
EuroEval leaderboard Not tested since Mar 2026

Source: EuroEval’s leaderboard

Scores over time

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

Plot with other benchmarks
On EuroEval’s leaderboard Best so far
0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) Mistral Small 3.2 Llama 3.1-70B Gemma 7B Llama 2-70B 0% 20% 40% 60% 80% 100% 2024 2025 2026 Near the ceiling (85%) Mistral Small 3.2 Gemma 7B

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

9 of 10 checks pass, 1 warning. Look first at the Newest score check. The newest came 8 months ago, in March 2026, for Nemotron 3 Nano 4B. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 1 warning 3 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (336 entries).
  • Pass Models scored 91 models we track, plus 245 other entries on its leaderboard.
  • Warning Newest score The newest came 8 months ago, in March 2026, for Nemotron 3 Nano 4B.
  • Note Frontier models None of the top 10 has a score. The strongest tested is Qwen3-235B-A22B (#131).
  • Pass Headroom The best is 79.9%, under 85%, so it still separates models.
Leaderboard import 3 pass
  • Pass Importer We read the board on EuroEval: 103 rows for models we track and 245 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