Benchmark Almanac

EuroEval

GPT-6.2 Astra 1.17 leads since November 2026
Not in model cards

Models answer questions, summarize text, and perform language understanding tasks across European languages. EuroEval combines task rankings on its European leaderboard.

The project began with Danish, Swedish, and Norwegian. The leaderboard also publishes results for individual languages.

Released
August 2021
Kind
Reading in many languages
Level
School
Scored
Checked automatically
Answers
Open-ended
Tasks
All public
Measures
Capability
Language
Many languages
In and out
Text → Text
Interaction
Single turn
Run by
Anyone
In views
Many languages
Strongest model tested
GPT-6.2 Astra #1 on the capability index
Last model added
November 2026
Scores refreshed
November 20th, 2026
143
Entries on its own leaderboard, 75 models we track

GPT-6.2 Astra leads with 1.17 GPT-6.2 Astra leads with 1.17

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

  1. 1 GPT-6.2 Astra zero-shot, val 1.17
  2. 2 GPT-6 Astra zero-shot, val 1.20
  3. 3 GPT-6.1 Sol zero-shot, val 1.21
  4. 4 GPT-6.2 Sol zero-shot, val 1.25
  5. 5 Gemini 3.7 Flash zero-shot, val 1.26
  6. 6 Gemini 3.6 Flash zero-shot, val 1.34
  7. 7 Gemini 3 Flash thinking, zero-shot, val 1.35
  8. 8 GLM-5.3-Flash val 1.38
  9. 9 GPT-5.6 Sol zero-shot, val 1.43
  10. 10 GPT-6 Sol zero-shot, val 1.48
  11. 11 deepseek/deepseek-flash low, zero-shot, val 1.54
  12. 12 Claude Sonnet 4.5 thinking, zero-shot, val 1.59
  13. 13 GPT-5.6 Terra zero-shot, val 1.61
  14. 14 Gemini 3.1 Flash-Lite zero-shot, val 1.62
  15. 15 GPT-5.4 Mini high, zero-shot, val 1.63
  16. 16 Qwen3.6 27B val 1.63
  17. 17 Qwen3.5-35B-A3B val 1.69
  18. 18 Qwen 3.8 27B val 1.70
  19. 19 Gemma 4 31B IT val 1.71
  20. 20 Grok 4.1 Fast zero-shot, val 1.72
  21. 21 Claude Sonnet 4.6 zero-shot, val 1.73
  22. 22 GPT-5 mini zero-shot, val 1.75
  23. 23 GPT-4.1 zero-shot, val 1.79
  24. 24 NbAiLab/borealis2-31b-preview val 1.80
  25. 25 GPT-5.4 none, zero-shot, val 1.80
  26. 26 Mistral Medium 3.5 val 1.80
  27. 27 Mistral Large 3 val 1.81
  28. 28 Gemini 3.5 Flash-Lite zero-shot, val 1.85
  29. 29 GPT-5.2 zero-shot, val 1.86
  30. 30 Gemini 2.5 Flash-Lite Sep 2025, thinking, zero-shot, val 1.91
  31. 31 GPT-5.6 Luna zero-shot, val 1.94
  32. 32 Qwen3.5-9B val 1.97
  33. 33 GPT-6 Luna zero-shot, val 1.99
  34. 34 GPT-5.4 Nano high, zero-shot, val 2.01
  35. 35 gpt-oss-120b val 2.05
  36. 36 typesafe/jev-1.13.0 zero-shot 2.06
  37. 37 Gemma 4 26B A4B val 2.07
  38. 38 Mistral Small 4 val 2.11
  39. 39 deepcogito/cogito-v1-preview-qwen-32B 2.13
  40. 40 Qwen3-32B no-thinking 2.13
  41. 41 GPT-5 nano zero-shot, val 2.15
  42. 42 Gemma 3 27B PT 2.16
  43. 43 Mistral Small 3.1 2.18
  44. 44 Mistral Small 3.2 2.18
  45. 45 NbAiLab/borealis2-26b-a4b-preview val 2.19
  46. 46 Gemma 4 12B val 2.19
  47. 47 Claude Haiku 4.5 zero-shot, val 2.23
  48. 48 mistralai/Mistral-Small-3.1-24B-Base-2503 2.24
  49. 49 mistralai/Ministral-3-14B-Instruct-2512 val 2.29
  50. 50 speakleash/Bielik-11B-v3.0-Instruct val 2.30
  51. 51 speakleash/Bielik-11B-v3.1-Instruct val 2.30
  52. 52 Qwen3.5-4B val 2.30
  53. 53 Apertus 1.5 70B val 2.30
  54. 54 Gemma 3 12B 2.33
  55. 55 Gemma 3 12B PT 2.37
  56. 56 Qwen3-14B no-thinking 2.37
  57. 57 deepcogito/cogito-v1-preview-qwen-14B 2.38
  58. 58 mistralai/Ministral-3-14B-Base-2512 val 2.38
  59. 59 Qwen3-4B-Thinking-2507 2.45
  60. 60 Ministral 3 14B val 2.45
  61. 61 speakleash/Bielik-Minitron-7B-v3.0-Instruct val 2.51
  62. 62 Nemotron 3 Nano 30B-A3B val 2.52
  63. 63 Ministral 3 8B Instruct val 2.53
  64. 64 mistralai/Ministral-3-8B-Base-2512 val 2.54
  65. 65 gpt-oss-20b low 2.57
  66. 66 Gemma 4 E4B val 2.57
  67. 67 speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c zero-shot 2.60
  68. 68 Qwen3-8B no-thinking 2.63
  69. 69 mistralai/Ministral-3-8B-Reasoning-2512 2.66
  70. 70 Apertus 1.5 8B val 2.67
  71. 71 Qwen3.5-9B-Base val 2.70
  72. 72 NbAiLab/borealis2-e4b-preview val 2.76
  73. 73 Qwen3-4B no-thinking 2.80
  74. 74 Ministral 3 3B Instruct val 2.80
  75. 75 deepcogito/cogito-v1-preview-llama-8B 2.81
  76. 76 mii-llm/nesso-4B 2.82
  77. 77 domyn/Domyn-Small-v1.0 val 2.85
  78. 78 Nemotron 3 Nano 30B-A3B Base val 2.86
  79. 79 amalia-llm/AMALIA-9B-0626-DPO val 2.87
  80. 80 Aleph-Alpha/Kolibri-1 val 2.88
  81. 81 speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551 zero-shot 2.89
  82. 82 meta-llama/Llama-3.1-8B-Instruct 2.91
  83. 83 amalia-llm/AMALIA-9B-0626-SFT val 2.93
  84. 84 allenai/Llama-3.1-Tulu-3-8B-SFT 2.98
  85. 85 EXAONE 4.0 32B val 2.99
  86. 86 danish-foundation-models/DFM-Mimir-v1.5 val 3.00
  87. 87 Qwen3.5-4B-Base val 3.00
  88. 88 Aya Expanse 8B val 3.07
  89. 89 Llama 3.1-8B 3.10
  90. 90 NbAiLab/borealis2-e2b-preview val 3.11
  91. 91 Qwen3.5-2B val 3.11
  92. 92 danish-foundation-models/DFM-Mimir val 3.15
  93. 93 AI-Sweden-Models/Llama-3-8B-instruct val 3.18
  94. 94 swiss-ai/Apertus-v1.1-4B val 3.21
  95. 95 swiss-ai/Apertus-v1.1-4B-Instruct val 3.22
  96. 96 Qwen3-1.7B no-thinking 3.23
  97. 97 deepcogito/cogito-v1-preview-llama-3B 3.30
  98. 98 Nemotron 3 Nano 4B 3.35
  99. 99 LiquidAI/LFM2.5-8B-A1B-Base val 3.37
  100. 100 BSC-LT/salamandra-7b-instruct 3.38
  101. 101 openeurollm/oellm-9b-128k-theta32m-v3 val 3.40
  102. 102 meta-llama/Llama-3.2-3B-Instruct 3.41
  103. 103 HuggingFaceTB/SmolLM3-3B-Base 3.43
  104. 104 OpenLLM-France/Luciole-8B-Instruct-1.1 val 3.45
  105. 105 occiglot/occiglot-7b-eu5-instruct 3.46
  106. 106 OmniJev/OneJev1.6-0.8B zero-shot 3.49
  107. 107 Qwen3.5-0.8B val 3.53
  108. 108 syvai/claes-v1 zero-shot 3.55
  109. 109 openGPT-X/Teuken-7B-instruct-v0.6 val 3.67
  110. 110 swiss-ai/Apertus-v1.1-1.5B val 3.67
  111. 111 Gemma 3 1B 3.69
  112. 112 occiglot/occiglot-7b-eu5 3.72
  113. 113 OpenLLM-France/Luciole-23B-Base val 3.73
  114. 114 CohereLabs/aya-23-8B val 3.77
  115. 115 swiss-ai/Apertus-v1.1-1.5B-Instruct val 3.80
  116. 116 BSC-LT/salamandra-7b 3.81
  117. 117 Gemma 3 1B PT 3.81
  118. 118 convaiinnovations/laya-multilingual zero-shot 3.84
  119. 119 OpenLLM-France/Luciole-8B-Base val 3.84
  120. 120 syvai/claes-preview zero-shot 3.84
  121. 121 Qwen3-0.6B no-thinking 3.84
  122. 122 Gemma 4 E2B val 3.88
  123. 123 fastino/GLiNER2.5-Decide zero-shot 3.89
  124. 124 Llama 3.2 1B 3.89
  125. 125 swiss-ai/Apertus-v1.1-0.5B val 3.92
  126. 126 Qwen3.5-0.8B-Base val 3.92
  127. 127 swiss-ai/Apertus-v1.1-0.5B-Instruct val 3.97
  128. 128 jaredpalmer/kev-0.5b zero-shot 4.00
  129. 129 LFM2-1.2B 4.00
  130. 130 OpenLLM-France/Luciole-1B-Instruct-1.1 val 4.06
  131. 131 PleIAs/Pleias-1.2b-Preview 4.06
  132. 132 inclusionAI/Ling-mini-2.0 val 4.09
  133. 133 LiquidAI/LFM2-700M 4.10
  134. 134 convaiinnovations/laya zero-shot 4.11
  135. 135 Gemma 3 270M 4.14
  136. 136 LiquidAI/LFM2-350M 4.17
  137. 137 meta-llama/Llama-3.2-1B-Instruct 4.17
  138. 138 OpenLLM-France/Luciole-1B-Base val 4.20
  139. 139 PleIAs/Pleias-350m-Preview 4.22
  140. 140 PleIAs/Monad 4.32
  141. 141 AIDC-AI/Marco-Mini-Global-Base val 4.34
  142. 142 Muse-research/Muse-3B val 4.35
  143. 143 openGPT-X/Teuken-7B-base-v0.6 val 4.35

Source: Official EuroEval leaderboard

  1. 1 GPT-6.2 Astra zero-shot, val 1.17
  2. 2 GPT-6 Astra zero-shot, val 1.20
  3. 3 GPT-6.1 Sol zero-shot, val 1.21
  4. 4 GPT-6.2 Sol zero-shot, val 1.25
  5. 5 Gemini 3.7 Flash zero-shot, val 1.26
  6. 6 Gemini 3.6 Flash zero-shot, val 1.34
  7. 7 Gemini 3 Flash thinking, zero-shot, val 1.35
  8. 8 GLM-5.3-Flash val 1.38
  9. 9 GPT-5.6 Sol zero-shot, val 1.43
  10. 10 GPT-6 Sol zero-shot, val 1.48
  11. 11 Claude Sonnet 4.5 thinking, zero-shot, val 1.59
  12. 12 GPT-5.6 Terra zero-shot, val 1.61
  13. 13 Gemini 3.1 Flash-Lite zero-shot, val 1.62
  14. 14 GPT-5.4 Mini high, zero-shot, val 1.63
  15. 15 Qwen3.6 27B val 1.63
  16. 16 Qwen3.5-35B-A3B val 1.69
  17. 17 Qwen 3.8 27B val 1.70
  18. 18 Gemma 4 31B IT val 1.71
  19. 19 Grok 4.1 Fast zero-shot, val 1.72
  20. 20 Claude Sonnet 4.6 zero-shot, val 1.73
  21. 21 GPT-5 mini zero-shot, val 1.75
  22. 22 GPT-4.1 zero-shot, val 1.79
  23. 23 GPT-5.4 none, zero-shot, val 1.80
  24. 24 Mistral Medium 3.5 val 1.80
  25. 25 Mistral Large 3 val 1.81
  26. 26 Gemini 3.5 Flash-Lite zero-shot, val 1.85
  27. 27 GPT-5.2 zero-shot, val 1.86
  28. 28 Gemini 2.5 Flash-Lite Sep 2025, thinking, zero-shot, val 1.91
  29. 29 GPT-5.6 Luna zero-shot, val 1.94
  30. 30 Qwen3.5-9B val 1.97
  31. 31 GPT-6 Luna zero-shot, val 1.99
  32. 32 GPT-5.4 Nano high, zero-shot, val 2.01
  33. 33 gpt-oss-120b val 2.05
  34. 34 Gemma 4 26B A4B val 2.07
  35. 35 Mistral Small 4 val 2.11
  36. 36 Qwen3-32B no-thinking 2.13
  37. 37 GPT-5 nano zero-shot, val 2.15
  38. 38 Gemma 3 27B PT 2.16
  39. 39 Mistral Small 3.1 2.18
  40. 40 Mistral Small 3.2 2.18
  41. 41 Gemma 4 12B val 2.19
  42. 42 Claude Haiku 4.5 zero-shot, val 2.23
  43. 43 Qwen3.5-4B val 2.30
  44. 44 Apertus 1.5 70B val 2.30
  45. 45 Gemma 3 12B 2.33
  46. 46 Gemma 3 12B PT 2.37
  47. 47 Qwen3-14B no-thinking 2.37
  48. 48 Qwen3-4B-Thinking-2507 2.45
  49. 49 Ministral 3 14B val 2.45
  50. 50 Nemotron 3 Nano 30B-A3B val 2.52
  51. 51 Ministral 3 8B Instruct val 2.53
  52. 52 gpt-oss-20b low 2.57
  53. 53 Gemma 4 E4B val 2.57
  54. 54 Qwen3-8B no-thinking 2.63
  55. 55 Apertus 1.5 8B val 2.67
  56. 56 Qwen3.5-9B-Base val 2.70
  57. 57 Qwen3-4B no-thinking 2.80
  58. 58 Ministral 3 3B Instruct val 2.80
  59. 59 Nemotron 3 Nano 30B-A3B Base val 2.86
  60. 60 EXAONE 4.0 32B val 2.99
  61. 61 Qwen3.5-4B-Base val 3.00
  62. 62 Aya Expanse 8B val 3.07
  63. 63 Llama 3.1-8B 3.10
  64. 64 Qwen3.5-2B val 3.11
  65. 65 Qwen3-1.7B no-thinking 3.23
  66. 66 Nemotron 3 Nano 4B 3.35
  67. 67 Qwen3.5-0.8B val 3.53
  68. 68 Gemma 3 1B 3.69
  69. 69 Gemma 3 1B PT 3.81
  70. 70 Qwen3-0.6B no-thinking 3.84
  71. 71 Gemma 4 E2B val 3.88
  72. 72 Llama 3.2 1B 3.89
  73. 73 Qwen3.5-0.8B-Base val 3.92
  74. 74 LFM2-1.2B 4.00
  75. 75 Gemma 3 270M 4.14

Source: Official EuroEval leaderboard

#EntryScore ±SettingDateLinkSource
1 GPT-6.2 Astra 1.17 ±0 zero-shot, val Nov 2026 platform.openai.com Official leaderboard
2 GPT-6 Astra 1.20 ±0.1 zero-shot, val Sep 2026 platform.openai.com Official leaderboard
3 GPT-6.1 Sol 1.21 ±0.1 zero-shot, val Sep 2026 platform.openai.com Official leaderboard
4 GPT-6.2 Sol 1.25 ±0.1 zero-shot, val Nov 2026 platform.openai.com Official leaderboard
5 Gemini 3.7 Flash 1.26 ±0.1 zero-shot, val Aug 2026 ai.google.dev Official leaderboard
6 Gemini 3.6 Flash 1.34 ±0.1 zero-shot, val Jul 2026 ai.google.dev Official leaderboard
7 Gemini 3 Flash 1.35 ±0 thinking, zero-shot, val Dec 2025 ai.google.dev Official leaderboard
8 GLM-5.3-Flash 1.38 ±0.1 val Aug 2026 hf.co Official leaderboard
9 GPT-5.6 Sol 1.43 ±0.1 zero-shot, val Jul 2026 platform.openai.com Official leaderboard
10 GPT-6 Sol 1.48 ±0.1 zero-shot, val Sep 2026 platform.openai.com Official leaderboard
11 deepseek/deepseek-flash closed weights, 552B 1.54 ±0.1 low, zero-shot, val api-docs.deepseek.com Official leaderboard
12 Claude Sonnet 4.5 1.59 ±0.1 thinking, zero-shot, val Sep 2025 docs.anthropic.com Official leaderboard
13 GPT-5.6 Terra 1.61 ±0.1 zero-shot, val Jul 2026 platform.openai.com Official leaderboard
14 Gemini 3.1 Flash-Lite 1.62 ±0.1 zero-shot, val Mar 2026 ai.google.dev Official leaderboard
15 GPT-5.4 Mini 1.63 ±0.1 high, zero-shot, val Mar 2026 platform.openai.com Official leaderboard
16 Qwen3.6 27B 1.63 ±0.1 val Apr 2026 hf.co Official leaderboard
17 Qwen3.5-35B-A3B 1.69 ±0.1 val Feb 2026 hf.co Official leaderboard
18 Qwen 3.8 27B 1.70 ±0.1 val Aug 2026 hf.co Official leaderboard
19 Gemma 4 31B IT 1.71 ±0.1 val Apr 2026 hf.co Official leaderboard
20 Grok 4.1 Fast 1.72 ±0.1 zero-shot, val Nov 2025 docs.x.ai Official leaderboard
21 Claude Sonnet 4.6 1.73 ±0.1 zero-shot, val Feb 2026 docs.anthropic.com Official leaderboard
22 GPT-5 mini 1.75 ±0.1 zero-shot, val Aug 2025 platform.openai.com Official leaderboard
23 GPT-4.1 1.79 ±0.1 zero-shot, val Apr 2025 platform.openai.com Official leaderboard
24 NbAiLab/borealis2-31b-preview open weights, 31.3B 1.80 ±0.1 val hf.co Official leaderboard
25 GPT-5.4 1.80 ±0.1 none, zero-shot, val Mar 2026 platform.openai.com Official leaderboard
26 Mistral Medium 3.5 1.80 ±0.1 val Apr 2026 hf.co Official leaderboard
27 Mistral Large 3 1.81 ±0.1 val Dec 2025 hf.co Official leaderboard
28 Gemini 3.5 Flash-Lite 1.85 ±0.1 zero-shot, val Jul 2026 ai.google.dev Official leaderboard
29 GPT-5.2 1.86 ±0.1 zero-shot, val Dec 2025 platform.openai.com Official leaderboard
30 Gemini 2.5 Flash-Lite Sep 2025 1.91 ±0.1 thinking, zero-shot, val Sep 2025 ai.google.dev Official leaderboard
31 GPT-5.6 Luna 1.94 ±0.1 zero-shot, val Jul 2026 platform.openai.com Official leaderboard
32 Qwen3.5-9B 1.97 ±0.1 val Feb 2026 hf.co Official leaderboard
33 GPT-6 Luna 1.99 ±0.1 zero-shot, val Sep 2026 platform.openai.com Official leaderboard
34 GPT-5.4 Nano 2.01 ±0.1 high, zero-shot, val Mar 2026 platform.openai.com Official leaderboard
35 gpt-oss-120b 2.05 ±0.1 val Aug 2025 hf.co Official leaderboard
36 typesafe/jev-1.13.0 closed weights 2.06 ±0.1 zero-shot docs.typesafe.ai Official leaderboard
37 Gemma 4 26B A4B 2.07 ±0.1 val Apr 2026 hf.co Official leaderboard
38 Mistral Small 4 2.11 ±0.1 val Mar 2026 hf.co Official leaderboard
39 deepcogito/cogito-v1-preview-qwen-32B open weights, 32.8B 2.13 ±0.1 hf.co Official leaderboard
40 Qwen3-32B 2.13 ±0.1 no-thinking Apr 2025 hf.co Official leaderboard
41 GPT-5 nano 2.15 ±0.1 zero-shot, val Aug 2025 platform.openai.com Official leaderboard
42 Gemma 3 27B PT 2.16 ±0.1 Mar 2025 hf.co Official leaderboard
43 Mistral Small 3.1 2.18 ±0.1 Mar 2025 hf.co Official leaderboard
44 Mistral Small 3.2 2.18 ±0.1 Jun 2025 hf.co Official leaderboard
45 NbAiLab/borealis2-26b-a4b-preview open weights, 25.8B 2.19 ±0.1 val hf.co Official leaderboard
46 Gemma 4 12B 2.19 ±0.1 val Jun 2026 hf.co Official leaderboard
47 Claude Haiku 4.5 2.23 ±0.1 zero-shot, val Oct 2025 docs.anthropic.com Official leaderboard
48 mistralai/Mistral-Small-3.1-24B-Base-2503 open weights, 24B 2.24 ±0.1 hf.co Official leaderboard
49 mistralai/Ministral-3-14B-Instruct-2512 open weights, 14B 2.29 ±0.1 val hf.co Official leaderboard
50 speakleash/Bielik-11B-v3.0-Instruct open weights, 11.2B 2.30 ±0.1 val hf.co Official leaderboard
51 speakleash/Bielik-11B-v3.1-Instruct open weights, 11.2B 2.30 ±0.1 val hf.co Official leaderboard
52 Qwen3.5-4B 2.30 ±0.1 val Feb 2026 hf.co Official leaderboard
53 Apertus 1.5 70B 2.30 ±0.1 val Jul 2026 hf.co Official leaderboard
54 Gemma 3 12B 2.33 ±0.1 Mar 2025 hf.co Official leaderboard
55 Gemma 3 12B PT 2.37 ±0.1 Mar 2025 hf.co Official leaderboard
56 Qwen3-14B 2.37 ±0.1 no-thinking Apr 2025 hf.co Official leaderboard
57 deepcogito/cogito-v1-preview-qwen-14B open weights, 14.8B 2.38 ±0.1 hf.co Official leaderboard
58 mistralai/Ministral-3-14B-Base-2512 open weights, 14B 2.38 ±0.1 val hf.co Official leaderboard
59 Qwen3-4B-Thinking-2507 2.45 ±0.1 Aug 2025 hf.co Official leaderboard
60 Ministral 3 14B 2.45 ±0.2 val Dec 2025 hf.co Official leaderboard
61 speakleash/Bielik-Minitron-7B-v3.0-Instruct open weights, 7.5B 2.51 ±0.1 val hf.co Official leaderboard
62 Nemotron 3 Nano 30B-A3B 2.52 ±0.1 val Dec 2025 hf.co Official leaderboard
63 Ministral 3 8B Instruct 2.53 ±0.1 val Dec 2025 hf.co Official leaderboard
64 mistralai/Ministral-3-8B-Base-2512 open weights, 8.9B 2.54 ±0.2 val hf.co Official leaderboard
65 gpt-oss-20b 2.57 ±0.1 low Aug 2025 hf.co Official leaderboard
66 Gemma 4 E4B 2.57 ±0.1 val Mar 2026 hf.co Official leaderboard
67 speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c open weights, 11.2B 2.60 ±0.1 zero-shot hf.co Official leaderboard
68 Qwen3-8B 2.63 ±0.1 no-thinking Apr 2025 hf.co Official leaderboard
69 mistralai/Ministral-3-8B-Reasoning-2512 open weights, 8.9B 2.66 ±0.1 hf.co Official leaderboard
70 Apertus 1.5 8B 2.67 ±0.1 val Jul 2026 hf.co Official leaderboard
71 Qwen3.5-9B-Base 2.70 ±0.2 val Feb 2026 hf.co Official leaderboard
72 NbAiLab/borealis2-e4b-preview open weights, 8B 2.76 ±0.1 val hf.co Official leaderboard
73 Qwen3-4B 2.80 ±0.1 no-thinking Apr 2025 hf.co Official leaderboard
74 Ministral 3 3B Instruct 2.80 ±0.1 val Dec 2025 hf.co Official leaderboard
75 deepcogito/cogito-v1-preview-llama-8B open weights, 8B 2.81 ±0.1 hf.co Official leaderboard
76 mii-llm/nesso-4B open weights, 4B 2.82 ±0.1 hf.co Official leaderboard
77 domyn/Domyn-Small-v1.0 open weights, 9.8B 2.85 ±0.1 val hf.co Official leaderboard
78 Nemotron 3 Nano 30B-A3B Base 2.86 ±0.2 val Dec 2025 hf.co Official leaderboard
79 amalia-llm/AMALIA-9B-0626-DPO open weights, 9.2B 2.87 ±0.1 val hf.co Official leaderboard
80 Aleph-Alpha/Kolibri-1 open weights, 78.1B 2.88 ±0.1 val hf.co Official leaderboard
81 speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551 open weights, 4.7B 2.89 ±0.2 zero-shot hf.co Official leaderboard
82 meta-llama/Llama-3.1-8B-Instruct open weights, 8B 2.91 ±0.2 hf.co Official leaderboard
83 amalia-llm/AMALIA-9B-0626-SFT open weights, 9.2B 2.93 ±0.1 val hf.co Official leaderboard
84 allenai/Llama-3.1-Tulu-3-8B-SFT open weights, 8B 2.98 ±0.1 hf.co Official leaderboard
85 EXAONE 4.0 32B 2.99 ±0.1 val Jul 2025 hf.co Official leaderboard
86 danish-foundation-models/DFM-Mimir-v1.5 open weights, 1.8B 3.00 ±0.1 val hf.co Official leaderboard
87 Qwen3.5-4B-Base 3.00 ±0.1 val Feb 2026 hf.co Official leaderboard
88 Aya Expanse 8B 3.07 ±0.1 val Oct 2024 hf.co Official leaderboard
89 Llama 3.1-8B 3.10 ±0.1 Jul 2024 hf.co Official leaderboard
90 NbAiLab/borealis2-e2b-preview open weights, 5.1B 3.11 ±0.1 val hf.co Official leaderboard
91 Qwen3.5-2B 3.11 ±0.1 val Feb 2026 hf.co Official leaderboard
92 danish-foundation-models/DFM-Mimir open weights, 1.8B 3.15 ±0.1 val hf.co Official leaderboard
93 AI-Sweden-Models/Llama-3-8B-instruct open weights, 8B 3.18 ±0.1 val hf.co Official leaderboard
94 swiss-ai/Apertus-v1.1-4B open weights, 3.8B 3.21 ±0.1 val hf.co Official leaderboard
95 swiss-ai/Apertus-v1.1-4B-Instruct open weights, 3.8B 3.22 ±0.1 val hf.co Official leaderboard
96 Qwen3-1.7B 3.23 ±0.1 no-thinking Apr 2025 hf.co Official leaderboard
97 deepcogito/cogito-v1-preview-llama-3B open weights, 3.6B 3.30 ±0.1 hf.co Official leaderboard
98 Nemotron 3 Nano 4B 3.35 ±0.1 Mar 2026 hf.co Official leaderboard
99 LiquidAI/LFM2.5-8B-A1B-Base open weights, 8.5B 3.37 ±0.1 val hf.co Official leaderboard
100 BSC-LT/salamandra-7b-instruct open weights, 7.8B 3.38 ±0.1 hf.co Official leaderboard
101 openeurollm/oellm-9b-128k-theta32m-v3 open weights 3.40 ±0.1 val hf.co Official leaderboard
102 meta-llama/Llama-3.2-3B-Instruct open weights, 3.2B 3.41 ±0.1 hf.co Official leaderboard
103 HuggingFaceTB/SmolLM3-3B-Base open weights, 3.1B 3.43 ±0.1 hf.co Official leaderboard
104 OpenLLM-France/Luciole-8B-Instruct-1.1 open weights, 8.1B 3.45 ±0.1 val hf.co Official leaderboard
105 occiglot/occiglot-7b-eu5-instruct open weights, 7.2B 3.46 ±0.1 hf.co Official leaderboard
106 OmniJev/OneJev1.6-0.8B open weights, 0.9B 3.49 ±0.2 zero-shot hf.co Official leaderboard
107 Qwen3.5-0.8B 3.53 ±0.1 val Feb 2026 hf.co Official leaderboard
108 syvai/claes-v1 open weights, 1.1B 3.55 ±0.1 zero-shot hf.co Official leaderboard
109 openGPT-X/Teuken-7B-instruct-v0.6 open weights, 7.5B 3.67 ±0.1 val hf.co Official leaderboard
110 swiss-ai/Apertus-v1.1-1.5B open weights, 1.5B 3.67 ±0.1 val hf.co Official leaderboard
111 Gemma 3 1B 3.69 ±0.1 Mar 2025 hf.co Official leaderboard
112 occiglot/occiglot-7b-eu5 open weights, 7.2B 3.72 ±0.1 hf.co Official leaderboard
113 OpenLLM-France/Luciole-23B-Base open weights, 23.2B 3.73 ±0.1 val hf.co Official leaderboard
114 CohereLabs/aya-23-8B open weights, 8B 3.77 ±0.1 val hf.co Official leaderboard
115 swiss-ai/Apertus-v1.1-1.5B-Instruct open weights, 1.5B 3.80 ±0.1 val hf.co Official leaderboard
116 BSC-LT/salamandra-7b open weights, 7.8B 3.81 ±0.1 hf.co Official leaderboard
117 Gemma 3 1B PT 3.81 ±0.1 Mar 2025 hf.co Official leaderboard
118 convaiinnovations/laya-multilingual open weights, 0.3B 3.84 ±0.1 zero-shot hf.co Official leaderboard
119 OpenLLM-France/Luciole-8B-Base open weights, 8.1B 3.84 ±0.1 val hf.co Official leaderboard
120 syvai/claes-preview open weights, 0.3B 3.84 ±0.1 zero-shot hf.co Official leaderboard
121 Qwen3-0.6B 3.84 ±0.1 no-thinking Apr 2025 hf.co Official leaderboard
122 Gemma 4 E2B 3.88 ±0.1 val Mar 2026 hf.co Official leaderboard
123 fastino/GLiNER2.5-Decide open weights, 0.5B 3.89 ±0.1 zero-shot hf.co Official leaderboard
124 Llama 3.2 1B 3.89 ±0.1 Sep 2024 hf.co Official leaderboard
125 swiss-ai/Apertus-v1.1-0.5B open weights, 0.4B 3.92 ±0.1 val hf.co Official leaderboard
126 Qwen3.5-0.8B-Base 3.92 ±0.1 val Feb 2026 hf.co Official leaderboard
127 swiss-ai/Apertus-v1.1-0.5B-Instruct open weights, 0.6B 3.97 ±0.1 val hf.co Official leaderboard
128 jaredpalmer/kev-0.5b open weights, 0.5B 4.00 ±0.1 zero-shot hf.co Official leaderboard
129 LFM2-1.2B 4.00 ±0.1 Jul 2025 hf.co Official leaderboard
130 OpenLLM-France/Luciole-1B-Instruct-1.1 open weights, 1.3B 4.06 ±0.1 val hf.co Official leaderboard
131 PleIAs/Pleias-1.2b-Preview open weights, 1.2B 4.06 ±0.1 hf.co Official leaderboard
132 inclusionAI/Ling-mini-2.0 open weights, 16.3B 4.09 ±0.1 val hf.co Official leaderboard
133 LiquidAI/LFM2-700M open weights, 0.7B 4.10 ±0.1 hf.co Official leaderboard
134 convaiinnovations/laya open weights, 0.4B 4.11 ±0.1 zero-shot hf.co Official leaderboard
135 Gemma 3 270M 4.14 ±0.1 Aug 2025 hf.co Official leaderboard
136 LiquidAI/LFM2-350M open weights, 0.4B 4.17 ±0 hf.co Official leaderboard
137 meta-llama/Llama-3.2-1B-Instruct open weights, 1.2B 4.17 ±0 hf.co Official leaderboard
138 OpenLLM-France/Luciole-1B-Base open weights, 1.3B 4.20 ±0.1 val hf.co Official leaderboard
139 PleIAs/Pleias-350m-Preview open weights, 0.4B 4.22 ±0.1 hf.co Official leaderboard
140 PleIAs/Monad open weights, 0.1B 4.32 ±0.1 hf.co Official leaderboard
141 AIDC-AI/Marco-Mini-Global-Base open weights, 17.4B 4.34 ±0.1 val hf.co Official leaderboard
142 Muse-research/Muse-3B open weights, 3B 4.35 ±0.1 val hf.co Official leaderboard
143 openGPT-X/Teuken-7B-base-v0.6 open weights, 7.5B 4.35 ±0.1 val hf.co Official leaderboard
#ModelScore ±SettingDateLinkSource
1 GPT-6.2 Astra 1.17 ±0 zero-shot, val Nov 2026 platform.openai.com Official leaderboard
2 GPT-6 Astra 1.20 ±0.1 zero-shot, val Sep 2026 platform.openai.com Official leaderboard
3 GPT-6.1 Sol 1.21 ±0.1 zero-shot, val Sep 2026 platform.openai.com Official leaderboard
4 GPT-6.2 Sol 1.25 ±0.1 zero-shot, val Nov 2026 platform.openai.com Official leaderboard
5 Gemini 3.7 Flash 1.26 ±0.1 zero-shot, val Aug 2026 ai.google.dev Official leaderboard
6 Gemini 3.6 Flash 1.34 ±0.1 zero-shot, val Jul 2026 ai.google.dev Official leaderboard
7 Gemini 3 Flash 1.35 ±0 thinking, zero-shot, val Dec 2025 ai.google.dev Official leaderboard
8 GLM-5.3-Flash 1.38 ±0.1 val Aug 2026 hf.co Official leaderboard
9 GPT-5.6 Sol 1.43 ±0.1 zero-shot, val Jul 2026 platform.openai.com Official leaderboard
10 GPT-6 Sol 1.48 ±0.1 zero-shot, val Sep 2026 platform.openai.com Official leaderboard
11 Claude Sonnet 4.5 1.59 ±0.1 thinking, zero-shot, val Sep 2025 docs.anthropic.com Official leaderboard
12 GPT-5.6 Terra 1.61 ±0.1 zero-shot, val Jul 2026 platform.openai.com Official leaderboard
13 Gemini 3.1 Flash-Lite 1.62 ±0.1 zero-shot, val Mar 2026 ai.google.dev Official leaderboard
14 GPT-5.4 Mini 1.63 ±0.1 high, zero-shot, val Mar 2026 platform.openai.com Official leaderboard
15 Qwen3.6 27B 1.63 ±0.1 val Apr 2026 hf.co Official leaderboard
16 Qwen3.5-35B-A3B 1.69 ±0.1 val Feb 2026 hf.co Official leaderboard
17 Qwen 3.8 27B 1.70 ±0.1 val Aug 2026 hf.co Official leaderboard
18 Gemma 4 31B IT 1.71 ±0.1 val Apr 2026 hf.co Official leaderboard
19 Grok 4.1 Fast 1.72 ±0.1 zero-shot, val Nov 2025 docs.x.ai Official leaderboard
20 Claude Sonnet 4.6 1.73 ±0.1 zero-shot, val Feb 2026 docs.anthropic.com Official leaderboard
21 GPT-5 mini 1.75 ±0.1 zero-shot, val Aug 2025 platform.openai.com Official leaderboard
22 GPT-4.1 1.79 ±0.1 zero-shot, val Apr 2025 platform.openai.com Official leaderboard
23 GPT-5.4 1.80 ±0.1 none, zero-shot, val Mar 2026 platform.openai.com Official leaderboard
24 Mistral Medium 3.5 1.80 ±0.1 val Apr 2026 hf.co Official leaderboard
25 Mistral Large 3 1.81 ±0.1 val Dec 2025 hf.co Official leaderboard
26 Gemini 3.5 Flash-Lite 1.85 ±0.1 zero-shot, val Jul 2026 ai.google.dev Official leaderboard
27 GPT-5.2 1.86 ±0.1 zero-shot, val Dec 2025 platform.openai.com Official leaderboard
28 Gemini 2.5 Flash-Lite Sep 2025 1.91 ±0.1 thinking, zero-shot, val Sep 2025 ai.google.dev Official leaderboard
29 GPT-5.6 Luna 1.94 ±0.1 zero-shot, val Jul 2026 platform.openai.com Official leaderboard
30 Qwen3.5-9B 1.97 ±0.1 val Feb 2026 hf.co Official leaderboard
31 GPT-6 Luna 1.99 ±0.1 zero-shot, val Sep 2026 platform.openai.com Official leaderboard
32 GPT-5.4 Nano 2.01 ±0.1 high, zero-shot, val Mar 2026 platform.openai.com Official leaderboard
33 gpt-oss-120b 2.05 ±0.1 val Aug 2025 hf.co Official leaderboard
34 Gemma 4 26B A4B 2.07 ±0.1 val Apr 2026 hf.co Official leaderboard
35 Mistral Small 4 2.11 ±0.1 val Mar 2026 hf.co Official leaderboard
36 Qwen3-32B 2.13 ±0.1 no-thinking Apr 2025 hf.co Official leaderboard
37 GPT-5 nano 2.15 ±0.1 zero-shot, val Aug 2025 platform.openai.com Official leaderboard
38 Gemma 3 27B PT 2.16 ±0.1 Mar 2025 hf.co Official leaderboard
39 Mistral Small 3.1 2.18 ±0.1 Mar 2025 hf.co Official leaderboard
40 Mistral Small 3.2 2.18 ±0.1 Jun 2025 hf.co Official leaderboard
41 Gemma 4 12B 2.19 ±0.1 val Jun 2026 hf.co Official leaderboard
42 Claude Haiku 4.5 2.23 ±0.1 zero-shot, val Oct 2025 docs.anthropic.com Official leaderboard
43 Qwen3.5-4B 2.30 ±0.1 val Feb 2026 hf.co Official leaderboard
44 Apertus 1.5 70B 2.30 ±0.1 val Jul 2026 hf.co Official leaderboard
45 Gemma 3 12B 2.33 ±0.1 Mar 2025 hf.co Official leaderboard
46 Gemma 3 12B PT 2.37 ±0.1 Mar 2025 hf.co Official leaderboard
47 Qwen3-14B 2.37 ±0.1 no-thinking Apr 2025 hf.co Official leaderboard
48 Qwen3-4B-Thinking-2507 2.45 ±0.1 Aug 2025 hf.co Official leaderboard
49 Ministral 3 14B 2.45 ±0.2 val Dec 2025 hf.co Official leaderboard
50 Nemotron 3 Nano 30B-A3B 2.52 ±0.1 val Dec 2025 hf.co Official leaderboard
51 Ministral 3 8B Instruct 2.53 ±0.1 val Dec 2025 hf.co Official leaderboard
52 gpt-oss-20b 2.57 ±0.1 low Aug 2025 hf.co Official leaderboard
53 Gemma 4 E4B 2.57 ±0.1 val Mar 2026 hf.co Official leaderboard
54 Qwen3-8B 2.63 ±0.1 no-thinking Apr 2025 hf.co Official leaderboard
55 Apertus 1.5 8B 2.67 ±0.1 val Jul 2026 hf.co Official leaderboard
56 Qwen3.5-9B-Base 2.70 ±0.2 val Feb 2026 hf.co Official leaderboard
57 Qwen3-4B 2.80 ±0.1 no-thinking Apr 2025 hf.co Official leaderboard
58 Ministral 3 3B Instruct 2.80 ±0.1 val Dec 2025 hf.co Official leaderboard
59 Nemotron 3 Nano 30B-A3B Base 2.86 ±0.2 val Dec 2025 hf.co Official leaderboard
60 EXAONE 4.0 32B 2.99 ±0.1 val Jul 2025 hf.co Official leaderboard
61 Qwen3.5-4B-Base 3.00 ±0.1 val Feb 2026 hf.co Official leaderboard
62 Aya Expanse 8B 3.07 ±0.1 val Oct 2024 hf.co Official leaderboard
63 Llama 3.1-8B 3.10 ±0.1 Jul 2024 hf.co Official leaderboard
64 Qwen3.5-2B 3.11 ±0.1 val Feb 2026 hf.co Official leaderboard
65 Qwen3-1.7B 3.23 ±0.1 no-thinking Apr 2025 hf.co Official leaderboard
66 Nemotron 3 Nano 4B 3.35 ±0.1 Mar 2026 hf.co Official leaderboard
67 Qwen3.5-0.8B 3.53 ±0.1 val Feb 2026 hf.co Official leaderboard
68 Gemma 3 1B 3.69 ±0.1 Mar 2025 hf.co Official leaderboard
69 Gemma 3 1B PT 3.81 ±0.1 Mar 2025 hf.co Official leaderboard
70 Qwen3-0.6B 3.84 ±0.1 no-thinking Apr 2025 hf.co Official leaderboard
71 Gemma 4 E2B 3.88 ±0.1 val Mar 2026 hf.co Official leaderboard
72 Llama 3.2 1B 3.89 ±0.1 Sep 2024 hf.co Official leaderboard
73 Qwen3.5-0.8B-Base 3.92 ±0.1 val Feb 2026 hf.co Official leaderboard
74 LFM2-1.2B 4.00 ±0.1 Jul 2025 hf.co Official leaderboard
75 Gemma 3 270M 4.14 ±0.1 Aug 2025 hf.co Official leaderboard

Rankings by part

We read 32 parts from the benchmark’s own leaderboard, each ranking models on one part of the test, such as Albanian, Belarusian and Bosnian. GPT-6.2 Astra leads 13 of them, and 7 other models lead the rest. Arrows show how far a model moved from its overall rank among the models we track.

Albanian

75 models we track have a score here.

  1. 1 GPT-6 Astra ↑1 1.37
  2. 2 Gemini 3.7 Flash ↑3 1.38
  3. 3 GPT-6.2 Astra ↓2 1.39
  4. 4 Gemini 3.6 Flash ↑2 1.45
  5. 5 GPT-6.1 Sol ↓2 1.48
  6. 6 GPT-6.2 Sol ↓2 1.49
  7. 7 GPT-5.6 Sol ↑2 1.60
  8. 8 GLM-5.3-Flash 1.62
  9. 9 Gemini 3 Flash ↓2 1.70
  10. 10 GPT-4.1 ↑12 1.73

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
Nemotron 3 Nano 4B
val
3.42
Ministral 3 3B Instruct
val zero-shot, …
3.24
Apertus 1.5 8B
val zero-shot, …
2.87
gpt-oss-20b
low
2.65
Apertus 1.5 70B
val zero-shot, …
2.61
Mistral Small 4
val zero-shot, …
2.48
GPT-5.4 Nano
none, zero-… low, zero-s…
2.82
Gemini 2.5 Flash-Lite
thinking, z… no-thinking…
2.35
Mistral Large 3
val zero-shot, …
2.18
Mistral Medium 3.5
val zero-shot, …
2.18
GPT-5.4 Mini
none, zero-… low, zero-s…
2.06
Claude Sonnet 4.5
thinking, z… no-thinking…
2.00

Source: Official EuroEval leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Gemma 3 270M 4.18
Gemma 3 270M 4.14
LFM2-1.2B 4.00
Qwen3.5-0.8B-Base val 3.92
Llama 3.2 1B 3.89
Gemma 4 E2B val 3.88
Qwen3-0.6B no-thinking 3.84
Gemma 3 1B PT 3.81
Gemma 3 1B 3.69
Gemma 4 E4B val 3.62
Qwen3.5-0.8B val 3.53
Nemotron 3 Nano 4B val 3.42
Nemotron 3 Nano 4B 3.35
Ministral 3 3B Instruct zero-shot, val 3.24
Qwen3-1.7B no-thinking 3.23
Qwen3.5-2B val 3.11
Llama 3.1-8B 3.10
Aya Expanse 8B val 3.07
Qwen3.5-4B-Base val 3.00
EXAONE 4.0 32B val 2.99
Gemma 4 26B A4B val 2.98
Gemma 4 12B val 2.87
Apertus 1.5 8B zero-shot, val 2.87
Nemotron 3 Nano 30B-A3B Base val 2.86
GPT-5.4 Nano none, zero-shot, val 2.82
Qwen3-4B no-thinking 2.80
Ministral 3 3B Instruct val 2.80
Qwen3.5-9B-Base val 2.70
Apertus 1.5 8B val 2.67
gpt-oss-20b high 2.65
Qwen3-8B no-thinking 2.63
Gemma 4 31B IT val 2.61
Apertus 1.5 70B zero-shot, val 2.61
Nemotron 3 Nano 30B-A3B val 2.60
gpt-oss-20b low 2.57
Gemma 4 E4B val 2.57
Ministral 3 8B Instruct val 2.53
Nemotron 3 Nano 30B-A3B val 2.52
Mistral Small 4 zero-shot, val 2.48
Qwen3-4B-Thinking-2507 2.45
Ministral 3 14B val 2.45
Gemma 3 12B PT 2.37
Qwen3-14B no-thinking 2.37
Gemini 2.5 Flash-Lite Sep 2025 no-thinking, zero-shot, val 2.35
Gemma 3 12B 2.33
Qwen3.5-4B val 2.30
Apertus 1.5 70B val 2.30
Claude Haiku 4.5 zero-shot, val 2.23
Gemma 4 12B val 2.19
Mistral Small 3.1 2.18
Mistral Small 3.2 2.18
Mistral Large 3 zero-shot, val 2.18
Mistral Medium 3.5 zero-shot, val 2.18
Gemma 3 27B PT 2.16
GPT-5 nano zero-shot, val 2.15
Qwen3-32B no-thinking 2.13
Gemini 3 Flash no-thinking, zero-shot, val 2.13
GPT-5.4 Nano low, zero-shot, val 2.12
Mistral Small 4 val 2.11
Gemma 4 26B A4B val 2.07
GPT-5.4 Mini none, zero-shot, val 2.06
gpt-oss-120b val 2.05
GPT-5.4 Nano high, zero-shot, val 2.01
GPT-5.4 Nano medium, zero-shot, val 2.01
Claude Sonnet 4.5 no-thinking, zero-shot, val 2.00
GPT-6 Luna zero-shot, val 1.99
Qwen3.5-9B val 1.97
GPT-5.6 Luna zero-shot, val 1.94
Gemini 2.5 Flash-Lite Sep 2025 thinking, zero-shot, val 1.91
GPT-5.2 zero-shot, val 1.86
Gemini 3.5 Flash-Lite zero-shot, val 1.85
Mistral Large 3 val 1.81
GPT-5.4 none, zero-shot, val 1.80
Mistral Medium 3.5 val 1.80
GPT-4.1 zero-shot, val 1.79
GPT-5 mini zero-shot, val 1.75
GPT-5.4 Mini low, zero-shot, val 1.74
Claude Sonnet 4.6 zero-shot, val 1.73
Grok 4.1 Fast zero-shot, val 1.72
Gemma 4 31B IT val 1.71
Qwen 3.8 27B val 1.70
Qwen3.5-35B-A3B val 1.69
GPT-5.4 Mini medium, zero-shot, val 1.68
GPT-5.4 Mini high, zero-shot, val 1.63
Qwen3.6 27B val 1.63
Gemini 3.1 Flash-Lite zero-shot, val 1.62
GPT-5.6 Terra zero-shot, val 1.61
Claude Sonnet 4.5 thinking, zero-shot, val 1.59
GLM-5.3-Flash zero-shot, val 1.56
GPT-6 Sol zero-shot, val 1.48
GPT-5.6 Sol zero-shot, val 1.43
GLM-5.3-Flash val 1.38
Gemini 3 Flash thinking, zero-shot, val 1.35
Gemini 3.6 Flash zero-shot, val 1.34
Gemini 3.7 Flash zero-shot, val 1.26
GPT-6.2 Sol zero-shot, val 1.25
GPT-6.1 Sol zero-shot, val 1.21
GPT-6 Astra zero-shot, val 1.20
GPT-6.2 Astra zero-shot, val 1.17

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
Nemotron 3 Nano 4B
val
3.42
Ministral 3 3B Instruct
val zero-shot, …
3.24
Apertus 1.5 8B
val zero-shot, …
2.87
gpt-oss-20b
low
2.65
Apertus 1.5 70B
val zero-shot, …
2.61
Mistral Small 4
val zero-shot, …
2.48
GPT-5.4 Nano
none, zero-… low, zero-s…
2.82
Gemini 2.5 Flash-Lite
thinking, z… no-thinking…
2.35
Mistral Large 3
val zero-shot, …
2.18
Mistral Medium 3.5
val zero-shot, …
2.18
GPT-5.4 Mini
none, zero-… low, zero-s…
2.06
Claude Sonnet 4.5
thinking, z… no-thinking…
2.00

Source: Official EuroEval leaderboard

ModelEffortHarnessScoreCost per taskOutput tokens
Gemma 3 270M 4.18
Gemma 3 270M 4.14
LFM2-1.2B 4.00
Qwen3.5-0.8B-Base val 3.92
Llama 3.2 1B 3.89
Gemma 4 E2B val 3.88
Qwen3-0.6B no-thinking 3.84
Gemma 3 1B PT 3.81
Gemma 3 1B 3.69
Gemma 4 E4B val 3.62
Qwen3.5-0.8B val 3.53
Nemotron 3 Nano 4B val 3.42
Nemotron 3 Nano 4B 3.35
Ministral 3 3B Instruct zero-shot, val 3.24
Qwen3-1.7B no-thinking 3.23
Qwen3.5-2B val 3.11
Llama 3.1-8B 3.10
Aya Expanse 8B val 3.07
Qwen3.5-4B-Base val 3.00
EXAONE 4.0 32B val 2.99
Gemma 4 26B A4B val 2.98
Gemma 4 12B val 2.87
Apertus 1.5 8B zero-shot, val 2.87
Nemotron 3 Nano 30B-A3B Base val 2.86
GPT-5.4 Nano none, zero-shot, val 2.82
Qwen3-4B no-thinking 2.80
Ministral 3 3B Instruct val 2.80
Qwen3.5-9B-Base val 2.70
Apertus 1.5 8B val 2.67
gpt-oss-20b high 2.65
Qwen3-8B no-thinking 2.63
Gemma 4 31B IT val 2.61
Apertus 1.5 70B zero-shot, val 2.61
Nemotron 3 Nano 30B-A3B val 2.60
gpt-oss-20b low 2.57
Gemma 4 E4B val 2.57
Ministral 3 8B Instruct val 2.53
Nemotron 3 Nano 30B-A3B val 2.52
Mistral Small 4 zero-shot, val 2.48
Qwen3-4B-Thinking-2507 2.45
Ministral 3 14B val 2.45
Gemma 3 12B PT 2.37
Qwen3-14B no-thinking 2.37
Gemini 2.5 Flash-Lite Sep 2025 no-thinking, zero-shot, val 2.35
Gemma 3 12B 2.33
Qwen3.5-4B val 2.30
Apertus 1.5 70B val 2.30
Claude Haiku 4.5 zero-shot, val 2.23
Gemma 4 12B val 2.19
Mistral Small 3.1 2.18
Mistral Small 3.2 2.18
Mistral Large 3 zero-shot, val 2.18
Mistral Medium 3.5 zero-shot, val 2.18
Gemma 3 27B PT 2.16
GPT-5 nano zero-shot, val 2.15
Qwen3-32B no-thinking 2.13
Gemini 3 Flash no-thinking, zero-shot, val 2.13
GPT-5.4 Nano low, zero-shot, val 2.12
Mistral Small 4 val 2.11
Gemma 4 26B A4B val 2.07
GPT-5.4 Mini none, zero-shot, val 2.06
gpt-oss-120b val 2.05
GPT-5.4 Nano high, zero-shot, val 2.01
GPT-5.4 Nano medium, zero-shot, val 2.01
Claude Sonnet 4.5 no-thinking, zero-shot, val 2.00
GPT-6 Luna zero-shot, val 1.99
Qwen3.5-9B val 1.97
GPT-5.6 Luna zero-shot, val 1.94
Gemini 2.5 Flash-Lite Sep 2025 thinking, zero-shot, val 1.91
GPT-5.2 zero-shot, val 1.86
Gemini 3.5 Flash-Lite zero-shot, val 1.85
Mistral Large 3 val 1.81
GPT-5.4 none, zero-shot, val 1.80
Mistral Medium 3.5 val 1.80
GPT-4.1 zero-shot, val 1.79
GPT-5 mini zero-shot, val 1.75
GPT-5.4 Mini low, zero-shot, val 1.74
Claude Sonnet 4.6 zero-shot, val 1.73
Grok 4.1 Fast zero-shot, val 1.72
Gemma 4 31B IT val 1.71
Qwen 3.8 27B val 1.70
Qwen3.5-35B-A3B val 1.69
GPT-5.4 Mini medium, zero-shot, val 1.68
GPT-5.4 Mini high, zero-shot, val 1.63
Qwen3.6 27B val 1.63
Gemini 3.1 Flash-Lite zero-shot, val 1.62
GPT-5.6 Terra zero-shot, val 1.61
Claude Sonnet 4.5 thinking, zero-shot, val 1.59
GLM-5.3-Flash zero-shot, val 1.56
GPT-6 Sol zero-shot, val 1.48
GPT-5.6 Sol zero-shot, val 1.43
GLM-5.3-Flash val 1.38
Gemini 3 Flash thinking, zero-shot, val 1.35
Gemini 3.6 Flash zero-shot, val 1.34
Gemini 3.7 Flash zero-shot, val 1.26
GPT-6.2 Sol zero-shot, val 1.25
GPT-6.1 Sol zero-shot, val 1.21
GPT-6 Astra zero-shot, val 1.20
GPT-6.2 Astra zero-shot, val 1.17

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

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

Meta Cohere Google DeepMind OpenAI Anthropic
Own leaderboard 11 leaders

Source: Official EuroEval leaderboard

Scores over time

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

Plot with other benchmarks
On the benchmark’s own leaderboard
1.00 2.00 3.00 4.00 5.00 2025 2026 2027 1.00 2.00 3.00 4.00 5.00 2025 2026 2027

Source: Official EuroEval leaderboard

No lab model card reports it yet

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

More in Multilingual

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

Data audit

All 10 checks pass. Every benchmark

Catalog record 2 pass
  • Pass Task access All tasks are public.
  • Pass Confidence Checked against a primary source.
Scores 4 pass 1 note
  • Pass Independent runs Scores from the official leaderboard (143 entries).
  • Pass Models scored 75 models we track, plus 68 other entries on its leaderboard.
  • Pass Newest score The newest is from November 2026, for GPT-6.2 Astra.
  • Pass Frontier models 3 of the top 10 have a score, led by GPT-6.2 Astra (#1).
  • Note Headroom Scored as a plain number, with no fixed ceiling.
Leaderboard import 3 pass
  • Pass Importer We read the board on euroeval.com: 99 rows for models we track and 68 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