Dutch CoLA
Models judge whether Dutch sentences drawn from descriptions of Dutch syntax are grammatically acceptable.
EuroEval compares the labels with human or constructed reference judgments using the Matthews correlation coefficient.
- Released
- April 2024
- Kind
- Reasoning in many languages
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Multiple choice
- Tasks
- All public
- Measures
- Capability
- Subject
- Languages and literature
- Language
- Dutch
- Interaction
- Single turn
- Items from
- Real work and use
- Refresh
- One fixed set
- Run by
- Anyone
- Strongest model tested
- gpt-oss-20b #143 on the capability index
- Last model added
- December 2025
- Scores refreshed
- November 21st, 2026
gemini/gemini-3.7-flash (zero-shot, val) tops the leaderboard with 64.8% Gemma 3 27B leads with 49.4% among models released by December 2025
Best score per entry from the benchmark's own leaderboard, read on November 21st, 2026. Where a model ran with several settings or harnesses, we keep its best. Names in gray aren't in our model list: fine-tunes, agent systems, products and older models. They don't count toward leaders. The best model we track is Gemma 3 27B, at 49.4%. This view shows only models we track, each at its best.
The newest model with a score here came out in December 2025, and nobody has tested a newer one. Among the 20 highest-ranked models on the capability index, these have no score yet:
- GPT-6.2 Astra
- Claude Fable 5.2
-
Gemini 4 Carbon
- GPT-6 Astra
- and 16 more
- 64.8 gemini/gemini-3.7-flash (zero-shot, val)
- 61.4 gemini/gemini-3.6-flash (zero-shot, val)
- 60.5 gpt-6.2-astra (zero-shot, val)
- 60.3 gpt-6.2-sol (zero-shot, val)
- 59.7 gemini/gemini-3.5-flash-lite (zero-shot, val)
- 58.6 mistralai/Mistral-Large-3-675B-Instruct-2512 (val)
- 57.8 openai/gpt-5.6-luna (zero-shot, val)
- 57.4 gemini/gemini-3-flash-preview#thinking (zero-shot, val)
- 57.3 NbAiLab/borealis2-31b-preview (val)
- 57.3 claude-sonnet-4-5-20250929#no-thinking (zero-shot, val)
- 56.4 gpt-4.1-2025-04-14 (zero-shot, val)
- 56.2 deepseek/deepseek-flash#no-thinking (val)
- 56.1 gpt-6.1-sol (zero-shot, val)
- 54.7 utter-project/EuroLLM-22B-Instruct-2512
- 54.6 zai-org/GLM-5.3-Flash (val)
- 54.3 gpt-6-astra (zero-shot, val)
- 54.2 gpt-6-sol (zero-shot, val)
- 54.1 gpt-5-mini-2025-08-07 (zero-shot, val)
- 54.0 gemini/gemini-3.1-flash-lite-preview (zero-shot, val)
- 53.7 mistralai/Mistral-Medium-3.5-128B (val)
- 53.0 google/gemma-4-31B-it (val)
- 52.9 claude-sonnet-4-5-20250929#thinking (zero-shot, val)
- 52.6 xai/grok-4-1-fast-reasoning (zero-shot, val)
- 52.2 mistralai/Mistral-Small-4-119B-2603 (val)
- 51.6 claude-sonnet-4-6 (zero-shot, val)
- 51.6 openai/gpt-5.4-mini-2026-03-17#none (zero-shot, val)
- 51.3 openai/gpt-5.4-mini-2026-03-17#low (zero-shot, val)
- 51.0 Qwen/Qwen3.6-27B-FP8 (val)
- 50.8 deepcogito/cogito-v1-preview-qwen-14B
- 50.6 mistralai/Mistral-Medium-3.5-128B (zero-shot, val)
- 50.6 openai/gpt-5.4-2026-03-05#none (zero-shot, val)
- 49.8 NbAiLab/borealis2-26b-a4b-preview (val)
- 49.7 speakleash/Bielik-11B-v3.1-Instruct (val)
- 49.6 mistralai/Ministral-3-14B-Base-2512 (val)
- 49.6 deepcogito/cogito-v2-preview-llama-70B
- 49.5 mistralai/Ministral-3-14B-Reasoning-2512 (val)
- 49.4
Gemma 3 27B
- 49.1 google/gemma-4-26B-A4B-it (val)
- 49.0 swiss-ai/Apertus-v1.5-70B (val)
- 48.9 Qwen/Qwen3.5-35B-A3B-FP8 (val)
- 48.8 gpt-6-luna (zero-shot, val)
- 48.7 NbAiLab/borealis2-26b-a4b-preview (zero-shot, val)
- 48.7 zai-org/GLM-5.3-Flash (zero-shot, val)
- 48.5
Gemma 3 12B PT
- 48.2
Gemma 3 27B PT
- 48.0 Qwen/Qwen3-14B#no-thinking
- 48.0 Qwen/Qwen3-32B#no-thinking
- 47.9 NbAiLab/borealis2-31b-preview (zero-shot, val)
- 47.7 gpt-5.6-sol (zero-shot, val)
- 47.5 deepcogito/cogito-v1-preview-qwen-32B
- 47.3 gemini/gemini-2.5-flash-lite#no-thinking (zero-shot, val)
- 46.6 amalia-llm/AMALIA-9B-0626-DPO (val)
- 46.4 deepseek/deepseek-flash#low (zero-shot, val)
- 46.3 gpt-5.2-2025-12-11 (zero-shot, val)
- 45.8
Gemma 3 12B
- 45.2 claude-haiku-4-5-20251001 (zero-shot, val)
- 45.1 gemini/gemini-2.5-flash-lite#thinking (zero-shot, val)
- 43.9 mistralai/Mistral-Small-3.1-24B-Base-2503
- 43.9 swiss-ai/Apertus-v1.5-8B (val)
- 43.7 openai/gpt-oss-120b (val)
- 43.1 speakleash/Bielik-11B-v3.0-Instruct (val)
- 43.1 openeurollm/prelude-long (val)
- 42.6 deepcogito/cogito-v1-preview-llama-8B
- 42.5 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 (val)
- 42.5 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-Base-BF16 (val)
- 42.2 latam-gpt/Llama-3.1-70B-LatamGPT-SFT-1.0 (val)
- 42.1 Qwen/Qwen3-8B#no-thinking
- 42.0 Qwen/Qwen3.5-9B (val)
- 41.9 gpt-5-nano-2025-08-07 (zero-shot, val)
- 41.6 openai/gpt-5.4-mini-2026-03-17#medium (zero-shot, val)
- 41.6 openai/gpt-5.4-nano-2026-03-17#medium (zero-shot, val)
- 41.3 allenai/Llama-3.1-Tulu-3-8B-SFT
- 41.3 google/gemma-4-12B-it (val)
- 41.2 mistralai/Mistral-Small-3.2-24B-Instruct-2506
- 40.9 openeurollm/prelude-base (val)
- 40.7 Qwen/Qwen3.8-27B (val)
- 40.2 gpt-5.6-terra (zero-shot, val)
- 39.9 speakleash/Bielik-Minitron-7B-v3.0-Instruct (val)
- 39.8 openai/gpt-5.4-nano-2026-03-17#low (zero-shot, val)
- 39.8 Qwen/Qwen3-4B#no-thinking
- 39.5 google/gemma-4-E4B-it (val)
- 39.3 mistralai/Mistral-Large-3-675B-Instruct-2512 (zero-shot, val)
- 39.1 speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c (zero-shot)
- 38.9 nvidia/Nemotron-Cascade-2-30B-A3B (val)
- 38.7 zai-org/GLM-4.7-Flash (val)
- 38.6 swiss-ai/Apertus-v1.5-70B (zero-shot, val)
- 38.1 domyn/Domyn-Small-v1.0 (val)
- 38.0 mistralai/Ministral-3-8B-Reasoning-2512
- 38.0
Llama 3.1-70B
- 37.8
Gemma 3 4B
- 37.6 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8 (val)
- 37.5 microsoft/mdeberta-v3-base
- 37.1 meta-llama/Llama-3.1-8B-Instruct
- 37.1 amalia-llm/AMALIA-9B-0626-SFT (val)
- 37.0 openai/gpt-5.4-nano-2026-03-17#high (zero-shot, val)
- 37.0 mistralai/Mistral-Small-3.1-24B-Instruct-2503
- 36.6 mistralai/Mistral-Small-4-119B-2603 (zero-shot, val)
- 36.3 gpt-oss-20b
- 36.3 Qwen/Qwen3.5-9B-Base (val)
- 36.3 mistralai/Ministral-3-14B-Instruct-2512 (val)
- 36.1 openai/gpt-5.4-mini-2026-03-17#high (zero-shot, val)
- 35.9 OpenLLM-France/Luciole-23B-Instruct-1.1 (val)
- 35.9 Qwen/Qwen3.5-35B-A3B-Base (val)
- 35.2 speakleash/Bielik-11B-v2.3-Instruct
- 35.2 NbAiLab/borealis2-e4b-preview (val)
- 34.7 CohereLabs/aya-expanse-8b (val)
- 33.9 utter-project/EuroLLM-9B-Instruct
- 33.6
Qwen3-4B-Thinking-2507
- 33.1 utter-project/EuroLLM-9B-Instruct-2512 (val)
- 32.4 OpenLLM-France/Luciole-8B-Instruct-1.1 (val)
- 32.2 google/gemma-4-31B (val)
- 31.6 Qwen/Qwen3.5-2B (val)
- 31.5 granite-4.0-micro
- 31.1 NbAiLab/borealis2-e2b-preview (val)
- 31.1
Gemma 3 4B PT
- 31.0 mii-llm/nesso-4B
- 30.7 allenai/Olmo-3-1125-32B
- 30.4 danish-foundation-models/DFM-Mimir-v1.8 (val)
- 29.8 Aleph-Alpha/Kolibri-1 (val)
- 29.8 google/gemma-4-E2B-it (val)
- 29.8 typesafe/jev-1.13.0 (zero-shot)
- 29.4 mistralai/Ministral-3-3B-Instruct-2512 (val)
- 29.0 deepcogito/cogito-v1-preview-llama-3B
- 28.8 allenai/Olmo-3.1-32B-Instruct-DPO
- 28.8 allenai/Olmo-3.1-32B-Instruct-SFT
- 28.5 danish-foundation-models/DFM-Mimir-v1.8 (zero-shot, val)
- 27.5
Olmo 3.1 32B Instruct
- 27.2 ibm-granite/granite-4.0-h-micro
- 27.1 Qwen/Qwen3.5-4B (val)
- 26.9
Ministral 3 3B Reasoning
- 26.8
Llama 3.1-8B
- 26.6 danish-foundation-models/DFM-Mimir-v1.5 (val)
- 25.9 danish-foundation-models/DFM-Mimir-v1.5 (zero-shot, val)
- 25.8 BSC-LT/salamandra-7b-instruct
- 25.5 amalia-llm/AMALIA-9B-0626-DPO (zero-shot, val)
- 24.8 amalia-llm/AMALIA-9B-0626-SFT (zero-shot, val)
- 24.6 swiss-ai/Apertus-v1.1-4B (val)
- 23.8 google/gemma-4-26B-A4B (val)
- 22.8 Qwen/Qwen3.5-2B-Base (val)
- 22.6 meta-llama/Llama-3.2-3B-Instruct
- 22.5 inclusionAI/Ling-mini-2.0 (val)
- 22.3 OpenLLM-France/Luciole-23B-Base (val)
- 22.0 google/gemma-4-E4B (val)
- 21.4 google/gemma-4-12B (val)
- 21.3 mistralai/Ministral-3-8B-Instruct-2512 (val)
- 21.2 Qwen/Qwen3.5-4B-Base (val)
- 20.5 LGAI-EXAONE/EXAONE-4.0-32B (val)
- 20.3 Qwen/Qwen3-1.7B#no-thinking
- 20.1 utter-project/EuroLLM-22B-2512
- 19.9 nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16
- 19.9 HuggingFaceTB/SmolLM3-3B-Base
- 19.8 nvidia/NVIDIA-Nemotron-3-Nano-4B-FP8 (val)
- 19.6 danish-foundation-models/DFM-Mimir (val)
- 19.3 Qwen/Qwen3.5-0.8B-Base (val)
- 19.0 openai/gpt-5.4-nano-2026-03-17#none (zero-shot, val)
- 18.5 jhu-clsp/mmBERT-small
- 18.4 jhu-clsp/mmBERT-base
- 18.1 swiss-ai/Apertus-v1.5-8B (zero-shot, val)
- 17.6 allenai/Olmo-3-1025-7B
- 17.3 speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551 (zero-shot)
- 17.1 danish-foundation-models/DFM-Mimir-v1.8-Mini (val)
- 16.7 JetBrains/Mellum2-12B-A2.5B-Instruct (val)
- 16.2 openGPT-X/Teuken-7B-instruct-v0.6 (val)
- 15.7 HuggingFaceTB/SmolLM3-3B#no-thinking
- 15.7 mistralai/Ministral-3-8B-Base-2512 (val)
- 15.5 occiglot/occiglot-7b-eu5-instruct
- 15.2 JetBrains/Mellum2-12B-A2.5B-Base (val)
- 14.9 Qwen/Qwen3.5-0.8B (val)
- 14.9 swiss-ai/Apertus-v1.1-4B-Instruct (val)
- 14.8 danish-foundation-models/DFM-Mimir-v1.8-Mini (zero-shot, val)
- 14.3 allenai/Olmo-3-7B-Instruct
- 13.9 openeurollm/oellm-9b-128k-theta32m-v3 (val)
- 13.6 allenai/Olmo-3-7B-Instruct-DPO
- 13.6 danish-foundation-models/DFM-Mimir (zero-shot, val)
- 12.6 LiquidAI/LFM2.5-8B-A1B (val)
- 12.4 intfloat/multilingual-e5-base
- 11.8 gemini/gemini-3-flash-preview#no-thinking (zero-shot, val)
- 11.4 intfloat/multilingual-e5-large
- 11.3 allenai/Olmo-3-7B-Instruct-SFT
- 11.3 EuroBERT/EuroBERT-210m
- 11.3 LiquidAI/LFM2.5-8B-A1B-Base (val)
- 11.2 answerdotai/ModernBERT-large
- 11.1 EuroBERT/EuroBERT-610m
- 9.84 CohereLabs/aya-23-8B (val)
- 9.65 Aleph-Alpha/Kolibri-1 (zero-shot, val)
- 9.52 speakleash/Bielik-1.5B-v3.0-Instruct
- 9.41 utter-project/EuroLLM-9B
- 8.80 TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49
- 7.68 distilbert/distilbert-base-multilingual-cased
- 7.56 Geotrend/distilbert-base-25lang-cased
- 7.55 google/gemma-4-E2B (val)
- 7.00
Qwen3-0.6B
- 6.56 answerdotai/ModernBERT-base
- 6.47 AI-Sweden-Models/bert-large-nordic-pile-1M-steps
- 6.40 swiss-ai/Apertus-v1.1-1.5B (val)
- 6.02 AI-Sweden-Models/Llama-3-8B-instruct (val)
- 5.99 OmniJev/OneJev1.6-0.8B (zero-shot)
- 5.90
Gemma 3 1B PT
- 5.74 EuropeanParliament/EUBERT
- 5.50 OpenLLM-France/Luciole-8B-Base (val)
- 5.00 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f (zero-shot)
- 4.90 TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894
- 4.70 fresh-electra-small
- 4.62 fresh-xlm-roberta-base
- 4.61 flwrlabs/Lizzy-7B (val)
- 4.37 NbAiLab/borealis2-e4b-preview (zero-shot, val)
- 4.30 syvai/claes-v1 (zero-shot)
- 4.00 convaiinnovations/laya (zero-shot)
- 3.88 occiglot/occiglot-7b-eu5
- 3.60 Qwen/Qwen3-0.6B#no-thinking
- 3.52 LFM2-1.2B
- 3.37 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base
- 3.34 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660 (zero-shot)
- 3.09 JetBrains/Mellum2-12B-A2.5B-Thinking-SFT (val)
- 2.98 PleIAs/Pleias-350m-Preview
- 2.90 yulan-team/YuLan-Mini-Instruct
- 2.78 meta-llama/Llama-3.2-1B-Instruct
- 2.62 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6
- 2.52
Gemma 3 1B
- 2.36
Gemma 3 270M
- 2.35
Llama 3.2 1B
- 1.95 BSC-LT/salamandra-7b
- 1.91 PleIAs/Pleias-1.2b-Preview
- 1.62 JetBrains/Mellum2-12B-A2.5B-Thinking (val)
- 1.08 LiquidAI/LFM2-350M
- 0.70
SmolLM2-135M
- 0.46
SmolLM2-360M
- 0.40 HuggingFaceTB/SmolLM2-135M-Instruct
- 0.23 google-bert/bert-base-multilingual-cased (val)
- 0.19 LiquidAI/LFM2-700M
- 0.14 jaredpalmer/kev-0.5b (zero-shot)
- 0.06 HuggingFaceTB/SmolLM2-360M-Instruct
- 0.0 AI-Sweden-Models/Llama-3-8B-instruct (zero-shot, val)
- 0.0 JetBrains/Mellum2-12B-A2.5B-Base-Pretrain (val)
- 0.0 mistralai/Ministral-3-14B-Instruct-2512 (zero-shot, val)
- 0.0 mistralai/Ministral-3-3B-Instruct-2512 (zero-shot, val)
- 0.0 Muse-research/Muse-3B (val)
- 0.0 openGPT-X/Teuken-7B-base-v0.6 (val)
- 0.0 PleIAs/Baguettotron@refs/pr/6#no-thinking
- 0.0 swiss-ai/Apertus-v1.1-1.5B-Instruct (val)
- 0.0 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview
- -0.06 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct
- -0.15 swiss-ai/Apertus-v1.1-1.5B (zero-shot)
- -0.23 NbAiLab/borealis2-e2b-preview (zero-shot, val)
- -0.32 convaiinnovations/laya-multilingual (zero-shot)
- -0.56 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660
- -0.92 JetBrains/Mellum2-12B-A2.5B-Instruct-SFT (val)
- -1.23 PleIAs/Monad
- -1.34 AIDC-AI/Marco-Mini-Global-Base (val)
- -1.76 syvai/claes-preview (zero-shot)
- -1.82 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6 (zero-shot)
- -2.26 OpenLLM-France/Luciole-1B-Base (val)
- -3.56 OpenLLM-France/Luciole-1B-Instruct-1.1 (val)
- -3.72 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f
- -3.89 swiss-ai/Apertus-v1.1-0.5B-Instruct (val)
- -5.92 fastino/GLiNER2.5-Decide (zero-shot)
- -6.66 swiss-ai/Apertus-v1.1-0.5B (val)
- 1 gemini/gemini-3.7-flash (zero-shot, val) 64.8%
- 2 gemini/gemini-3.6-flash (zero-shot, val) 61.4%
- 3 gpt-6.2-astra (zero-shot, val) 60.5%
- 4 gpt-6.2-sol (zero-shot, val) 60.3%
- 5 gemini/gemini-3.5-flash-lite (zero-shot, val) 59.7%
- 6 mistralai/Mistral-Large-3-675B-Instruct-2512 (val) 58.6%
- 7 openai/gpt-5.6-luna (zero-shot, val) 57.8%
- 8 gemini/gemini-3-flash-preview#thinking (zero-shot, val) 57.4%
- 9 NbAiLab/borealis2-31b-preview (val) 57.3%
- 10 claude-sonnet-4-5-20250929#no-thinking (zero-shot, val) 57.3%
- 11 gpt-4.1-2025-04-14 (zero-shot, val) 56.4%
- 12 deepseek/deepseek-flash#no-thinking (val) 56.2%
- 13 gpt-6.1-sol (zero-shot, val) 56.1%
- 14 utter-project/EuroLLM-22B-Instruct-2512 54.7%
- 15 zai-org/GLM-5.3-Flash (val) 54.6%
- 16 gpt-6-astra (zero-shot, val) 54.3%
- 17 gpt-6-sol (zero-shot, val) 54.2%
- 18 gpt-5-mini-2025-08-07 (zero-shot, val) 54.1%
- 19 gemini/gemini-3.1-flash-lite-preview (zero-shot, val) 54%
- 20 mistralai/Mistral-Medium-3.5-128B (val) 53.7%
- 21 google/gemma-4-31B-it (val) 53%
- 22 claude-sonnet-4-5-20250929#thinking (zero-shot, val) 52.9%
- 23 xai/grok-4-1-fast-reasoning (zero-shot, val) 52.6%
- 24 mistralai/Mistral-Small-4-119B-2603 (val) 52.2%
- 25 claude-sonnet-4-6 (zero-shot, val) 51.6%
- 26 openai/gpt-5.4-mini-2026-03-17#none (zero-shot, val) 51.6%
- 27 openai/gpt-5.4-mini-2026-03-17#low (zero-shot, val) 51.3%
- 28 Qwen/Qwen3.6-27B-FP8 (val) 51%
- 29 deepcogito/cogito-v1-preview-qwen-14B 50.8%
- 30 mistralai/Mistral-Medium-3.5-128B (zero-shot, val) 50.6%
- 31 openai/gpt-5.4-2026-03-05#none (zero-shot, val) 50.6%
- 32 NbAiLab/borealis2-26b-a4b-preview (val) 49.8%
- 33 speakleash/Bielik-11B-v3.1-Instruct (val) 49.7%
- 34 mistralai/Ministral-3-14B-Base-2512 (val) 49.6%
- 35 deepcogito/cogito-v2-preview-llama-70B 49.6%
- 36 mistralai/Ministral-3-14B-Reasoning-2512 (val) 49.5%
- 37
Gemma 3 27B 49.4%
- 38 google/gemma-4-26B-A4B-it (val) 49.1%
- 39 swiss-ai/Apertus-v1.5-70B (val) 49%
- 40 Qwen/Qwen3.5-35B-A3B-FP8 (val) 48.9%
- 41 gpt-6-luna (zero-shot, val) 48.8%
- 42 NbAiLab/borealis2-26b-a4b-preview (zero-shot, val) 48.7%
- 43 zai-org/GLM-5.3-Flash (zero-shot, val) 48.7%
- 44
Gemma 3 12B PT 48.5%
- 45
Gemma 3 27B PT 48.2%
- 46 Qwen/Qwen3-14B#no-thinking 48%
- 47 Qwen/Qwen3-32B#no-thinking 48%
- 48 NbAiLab/borealis2-31b-preview (zero-shot, val) 47.9%
- 49 gpt-5.6-sol (zero-shot, val) 47.7%
- 50 deepcogito/cogito-v1-preview-qwen-32B 47.5%
- 51 gemini/gemini-2.5-flash-lite#no-thinking (zero-shot, val) 47.3%
- 52 amalia-llm/AMALIA-9B-0626-DPO (val) 46.6%
- 53 deepseek/deepseek-flash#low (zero-shot, val) 46.4%
- 54 gpt-5.2-2025-12-11 (zero-shot, val) 46.3%
- 55
Gemma 3 12B 45.8%
- 56 claude-haiku-4-5-20251001 (zero-shot, val) 45.2%
- 57 gemini/gemini-2.5-flash-lite#thinking (zero-shot, val) 45.1%
- 58 mistralai/Mistral-Small-3.1-24B-Base-2503 43.9%
- 59 swiss-ai/Apertus-v1.5-8B (val) 43.9%
- 60 openai/gpt-oss-120b (val) 43.7%
- 61 speakleash/Bielik-11B-v3.0-Instruct (val) 43.1%
- 62 openeurollm/prelude-long (val) 43.1%
- 63 deepcogito/cogito-v1-preview-llama-8B 42.6%
- 64 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 (val) 42.5%
- 65 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-Base-BF16 (val) 42.5%
- 66 latam-gpt/Llama-3.1-70B-LatamGPT-SFT-1.0 (val) 42.2%
- 67 Qwen/Qwen3-8B#no-thinking 42.1%
- 68 Qwen/Qwen3.5-9B (val) 42%
- 69 gpt-5-nano-2025-08-07 (zero-shot, val) 41.9%
- 70 openai/gpt-5.4-mini-2026-03-17#medium (zero-shot, val) 41.6%
- 71 openai/gpt-5.4-nano-2026-03-17#medium (zero-shot, val) 41.6%
- 72 allenai/Llama-3.1-Tulu-3-8B-SFT 41.3%
- 73 google/gemma-4-12B-it (val) 41.3%
- 74 mistralai/Mistral-Small-3.2-24B-Instruct-2506 41.2%
- 75 openeurollm/prelude-base (val) 40.9%
- 76 Qwen/Qwen3.8-27B (val) 40.7%
- 77 gpt-5.6-terra (zero-shot, val) 40.2%
- 78 speakleash/Bielik-Minitron-7B-v3.0-Instruct (val) 39.9%
- 79 openai/gpt-5.4-nano-2026-03-17#low (zero-shot, val) 39.8%
- 80 Qwen/Qwen3-4B#no-thinking 39.8%
- 81 google/gemma-4-E4B-it (val) 39.5%
- 82 mistralai/Mistral-Large-3-675B-Instruct-2512 (zero-shot, val) 39.3%
- 83 speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c (zero-shot) 39.1%
- 84 nvidia/Nemotron-Cascade-2-30B-A3B (val) 38.9%
- 85 zai-org/GLM-4.7-Flash (val) 38.7%
- 86 swiss-ai/Apertus-v1.5-70B (zero-shot, val) 38.6%
- 87 domyn/Domyn-Small-v1.0 (val) 38.1%
- 88 mistralai/Ministral-3-8B-Reasoning-2512 38%
- 89
Llama 3.1-70B 38%
- 90
Gemma 3 4B 37.8%
- 91 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8 (val) 37.6%
- 92 microsoft/mdeberta-v3-base 37.5%
- 93 meta-llama/Llama-3.1-8B-Instruct 37.1%
- 94 amalia-llm/AMALIA-9B-0626-SFT (val) 37.1%
- 95 openai/gpt-5.4-nano-2026-03-17#high (zero-shot, val) 37%
- 96 mistralai/Mistral-Small-3.1-24B-Instruct-2503 37%
- 97 mistralai/Mistral-Small-4-119B-2603 (zero-shot, val) 36.6%
- 98 gpt-oss-20b low 36.3%
- 99 Qwen/Qwen3.5-9B-Base (val) 36.3%
- 100 mistralai/Ministral-3-14B-Instruct-2512 (val) 36.3%
- 101 openai/gpt-5.4-mini-2026-03-17#high (zero-shot, val) 36.1%
- 102 OpenLLM-France/Luciole-23B-Instruct-1.1 (val) 35.9%
- 103 Qwen/Qwen3.5-35B-A3B-Base (val) 35.9%
- 104 speakleash/Bielik-11B-v2.3-Instruct 35.2%
- 105 NbAiLab/borealis2-e4b-preview (val) 35.2%
- 106 CohereLabs/aya-expanse-8b (val) 34.7%
- 107 utter-project/EuroLLM-9B-Instruct 33.9%
- 108
Qwen3-4B-Thinking-2507 33.6%
- 109 utter-project/EuroLLM-9B-Instruct-2512 (val) 33.1%
- 110 OpenLLM-France/Luciole-8B-Instruct-1.1 (val) 32.4%
- 111 google/gemma-4-31B (val) 32.2%
- 112 Qwen/Qwen3.5-2B (val) 31.6%
- 113 granite-4.0-micro 31.5%
- 114 NbAiLab/borealis2-e2b-preview (val) 31.1%
- 115
Gemma 3 4B PT 31.1%
- 116 mii-llm/nesso-4B 31%
- 117 allenai/Olmo-3-1125-32B 30.7%
- 118 danish-foundation-models/DFM-Mimir-v1.8 (val) 30.4%
- 119 Aleph-Alpha/Kolibri-1 (val) 29.8%
- 120 google/gemma-4-E2B-it (val) 29.8%
- 121 typesafe/jev-1.13.0 (zero-shot) 29.8%
- 122 mistralai/Ministral-3-3B-Instruct-2512 (val) 29.4%
- 123 deepcogito/cogito-v1-preview-llama-3B 29%
- 124 allenai/Olmo-3.1-32B-Instruct-DPO 28.8%
- 125 allenai/Olmo-3.1-32B-Instruct-SFT 28.8%
- 126 danish-foundation-models/DFM-Mimir-v1.8 (zero-shot, val) 28.5%
- 127
Olmo 3.1 32B Instruct 27.5%
- 128 ibm-granite/granite-4.0-h-micro 27.2%
- 129 Qwen/Qwen3.5-4B (val) 27.1%
- 130
Ministral 3 3B Reasoning 26.9%
- 131
Llama 3.1-8B 26.8%
- 132 danish-foundation-models/DFM-Mimir-v1.5 (val) 26.6%
- 133 danish-foundation-models/DFM-Mimir-v1.5 (zero-shot, val) 25.9%
- 134 BSC-LT/salamandra-7b-instruct 25.8%
- 135 amalia-llm/AMALIA-9B-0626-DPO (zero-shot, val) 25.5%
- 136 amalia-llm/AMALIA-9B-0626-SFT (zero-shot, val) 24.8%
- 137 swiss-ai/Apertus-v1.1-4B (val) 24.6%
- 138 google/gemma-4-26B-A4B (val) 23.8%
- 139 Qwen/Qwen3.5-2B-Base (val) 22.8%
- 140 meta-llama/Llama-3.2-3B-Instruct 22.6%
- 141 inclusionAI/Ling-mini-2.0 (val) 22.5%
- 142 OpenLLM-France/Luciole-23B-Base (val) 22.3%
- 143 google/gemma-4-E4B (val) 22%
- 144 google/gemma-4-12B (val) 21.4%
- 145 mistralai/Ministral-3-8B-Instruct-2512 (val) 21.3%
- 146 Qwen/Qwen3.5-4B-Base (val) 21.2%
- 147 LGAI-EXAONE/EXAONE-4.0-32B (val) 20.5%
- 148 Qwen/Qwen3-1.7B#no-thinking 20.3%
- 149 utter-project/EuroLLM-22B-2512 20.1%
- 150 nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16 19.9%
- 151 HuggingFaceTB/SmolLM3-3B-Base 19.9%
- 152 nvidia/NVIDIA-Nemotron-3-Nano-4B-FP8 (val) 19.8%
- 153 danish-foundation-models/DFM-Mimir (val) 19.6%
- 154 Qwen/Qwen3.5-0.8B-Base (val) 19.3%
- 155 openai/gpt-5.4-nano-2026-03-17#none (zero-shot, val) 19%
- 156 jhu-clsp/mmBERT-small 18.5%
- 157 jhu-clsp/mmBERT-base 18.4%
- 158 swiss-ai/Apertus-v1.5-8B (zero-shot, val) 18.1%
- 159 allenai/Olmo-3-1025-7B 17.6%
- 160 speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551 (zero-shot) 17.3%
- 161 danish-foundation-models/DFM-Mimir-v1.8-Mini (val) 17.1%
- 162 JetBrains/Mellum2-12B-A2.5B-Instruct (val) 16.7%
- 163 openGPT-X/Teuken-7B-instruct-v0.6 (val) 16.2%
- 164 HuggingFaceTB/SmolLM3-3B#no-thinking 15.7%
- 165 mistralai/Ministral-3-8B-Base-2512 (val) 15.7%
- 166 occiglot/occiglot-7b-eu5-instruct 15.5%
- 167 JetBrains/Mellum2-12B-A2.5B-Base (val) 15.2%
- 168 Qwen/Qwen3.5-0.8B (val) 14.9%
- 169 swiss-ai/Apertus-v1.1-4B-Instruct (val) 14.9%
- 170 danish-foundation-models/DFM-Mimir-v1.8-Mini (zero-shot, val) 14.8%
- 171 allenai/Olmo-3-7B-Instruct 14.3%
- 172 openeurollm/oellm-9b-128k-theta32m-v3 (val) 13.9%
- 173 allenai/Olmo-3-7B-Instruct-DPO 13.6%
- 174 danish-foundation-models/DFM-Mimir (zero-shot, val) 13.6%
- 175 LiquidAI/LFM2.5-8B-A1B (val) 12.6%
- 176 intfloat/multilingual-e5-base 12.4%
- 177 gemini/gemini-3-flash-preview#no-thinking (zero-shot, val) 11.8%
- 178 intfloat/multilingual-e5-large 11.4%
- 179 allenai/Olmo-3-7B-Instruct-SFT 11.3%
- 180 EuroBERT/EuroBERT-210m 11.3%
- 181 LiquidAI/LFM2.5-8B-A1B-Base (val) 11.3%
- 182 answerdotai/ModernBERT-large 11.2%
- 183 EuroBERT/EuroBERT-610m 11.1%
- 184 CohereLabs/aya-23-8B (val) 9.84%
- 185 Aleph-Alpha/Kolibri-1 (zero-shot, val) 9.65%
- 186 speakleash/Bielik-1.5B-v3.0-Instruct 9.52%
- 187 utter-project/EuroLLM-9B 9.41%
- 188 TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49 8.80%
- 189 distilbert/distilbert-base-multilingual-cased 7.68%
- 190 Geotrend/distilbert-base-25lang-cased 7.56%
- 191 google/gemma-4-E2B (val) 7.55%
- 192
Qwen3-0.6B 7.00%
- 193 answerdotai/ModernBERT-base 6.56%
- 194 AI-Sweden-Models/bert-large-nordic-pile-1M-steps 6.47%
- 195 swiss-ai/Apertus-v1.1-1.5B (val) 6.40%
- 196 AI-Sweden-Models/Llama-3-8B-instruct (val) 6.02%
- 197 OmniJev/OneJev1.6-0.8B (zero-shot) 5.99%
- 198
Gemma 3 1B PT 5.90%
- 199 EuropeanParliament/EUBERT 5.74%
- 200 OpenLLM-France/Luciole-8B-Base (val) 5.50%
- 201 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f (zero-shot) 5.00%
- 202 TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894 4.90%
- 203 fresh-electra-small 4.70%
- 204 fresh-xlm-roberta-base 4.62%
- 205 flwrlabs/Lizzy-7B (val) 4.61%
- 206 NbAiLab/borealis2-e4b-preview (zero-shot, val) 4.37%
- 207 syvai/claes-v1 (zero-shot) 4.30%
- 208 convaiinnovations/laya (zero-shot) 4.00%
- 209 occiglot/occiglot-7b-eu5 3.88%
- 210 Qwen/Qwen3-0.6B#no-thinking 3.60%
- 211 LFM2-1.2B 3.52%
- 212 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base 3.37%
- 213 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660 (zero-shot) 3.34%
- 214 JetBrains/Mellum2-12B-A2.5B-Thinking-SFT (val) 3.09%
- 215 PleIAs/Pleias-350m-Preview 2.98%
- 216 yulan-team/YuLan-Mini-Instruct 2.90%
- 217 meta-llama/Llama-3.2-1B-Instruct 2.78%
- 218 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6 2.62%
- 219
Gemma 3 1B 2.52%
- 220
Gemma 3 270M 2.36%
- 221
Llama 3.2 1B 2.35%
- 222 BSC-LT/salamandra-7b 1.95%
- 223 PleIAs/Pleias-1.2b-Preview 1.91%
- 224 JetBrains/Mellum2-12B-A2.5B-Thinking (val) 1.62%
- 225 LiquidAI/LFM2-350M 1.08%
- 226
SmolLM2-135M 0.70%
- 227
SmolLM2-360M 0.46%
- 228 HuggingFaceTB/SmolLM2-135M-Instruct 0.40%
- 229 google-bert/bert-base-multilingual-cased (val) 0.23%
- 230 LiquidAI/LFM2-700M 0.19%
- 231 jaredpalmer/kev-0.5b (zero-shot) 0.14%
- 232 HuggingFaceTB/SmolLM2-360M-Instruct 0.06%
- 233 AI-Sweden-Models/Llama-3-8B-instruct (zero-shot, val) 0%
- 234 JetBrains/Mellum2-12B-A2.5B-Base-Pretrain (val) 0%
- 235 mistralai/Ministral-3-14B-Instruct-2512 (zero-shot, val) 0%
- 236 mistralai/Ministral-3-3B-Instruct-2512 (zero-shot, val) 0%
- 237 Muse-research/Muse-3B (val) 0%
- 238 openGPT-X/Teuken-7B-base-v0.6 (val) 0%
- 239 PleIAs/Baguettotron@refs/pr/6#no-thinking 0%
- 240 swiss-ai/Apertus-v1.1-1.5B-Instruct (val) 0%
- 241 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview 0%
- 242 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct -0.06%
- 243 swiss-ai/Apertus-v1.1-1.5B (zero-shot) -0.15%
- 244 NbAiLab/borealis2-e2b-preview (zero-shot, val) -0.23%
- 245 convaiinnovations/laya-multilingual (zero-shot) -0.32%
- 246 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660 -0.56%
- 247 JetBrains/Mellum2-12B-A2.5B-Instruct-SFT (val) -0.92%
- 248 PleIAs/Monad -1.23%
- 249 AIDC-AI/Marco-Mini-Global-Base (val) -1.34%
- 250 syvai/claes-preview (zero-shot) -1.76%
- 251 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6 (zero-shot) -1.82%
- 252 OpenLLM-France/Luciole-1B-Base (val) -2.26%
- 253 OpenLLM-France/Luciole-1B-Instruct-1.1 (val) -3.56%
- 254 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f -3.72%
- 255 swiss-ai/Apertus-v1.1-0.5B-Instruct (val) -3.89%
- 256 fastino/GLiNER2.5-Decide (zero-shot) -5.92%
- 257 swiss-ai/Apertus-v1.1-0.5B (val) -6.66%
Source: Official Dutch CoLA leaderboard
- 49.4
Gemma 3 27B
- 48.5
Gemma 3 12B PT
- 48.2
Gemma 3 27B PT
- 45.8
Gemma 3 12B
- 38.0
Llama 3.1-70B
- 37.8
Gemma 3 4B
- 36.3 gpt-oss-20b
- 33.6
Qwen3-4B-Thinking-2507
- 31.5 granite-4.0-micro
- 31.1
Gemma 3 4B PT
- 27.5
Olmo 3.1 32B Instruct
- 26.9
Ministral 3 3B Reasoning
- 26.8
Llama 3.1-8B
- 7.00
Qwen3-0.6B
- 5.90
Gemma 3 1B PT
- 3.52 LFM2-1.2B
- 2.52
Gemma 3 1B
- 2.36
Gemma 3 270M
- 2.35
Llama 3.2 1B
- 0.70
SmolLM2-135M
- 0.46
SmolLM2-360M
- 1
Gemma 3 27B 49.4%
- 2
Gemma 3 12B PT 48.5%
- 3
Gemma 3 27B PT 48.2%
- 4
Gemma 3 12B 45.8%
- 5
Llama 3.1-70B 38%
- 6
Gemma 3 4B 37.8%
- 7 gpt-oss-20b low 36.3%
- 8
Qwen3-4B-Thinking-2507 33.6%
- 9 granite-4.0-micro 31.5%
- 10
Gemma 3 4B PT 31.1%
- 11
Olmo 3.1 32B Instruct 27.5%
- 12
Ministral 3 3B Reasoning 26.9%
- 13
Llama 3.1-8B 26.8%
- 14
Qwen3-0.6B 7.00%
- 15
Gemma 3 1B PT 5.90%
- 16 LFM2-1.2B 3.52%
- 17
Gemma 3 1B 2.52%
- 18
Gemma 3 270M 2.36%
- 19
Llama 3.2 1B 2.35%
- 20
SmolLM2-135M 0.70%
- 21
SmolLM2-360M 0.46%
Source: Official Dutch CoLA leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | gemini/gemini-3.7-flash (zero-shot, val) | 64.8% | Official leaderboard | ||
| 2 | gemini/gemini-3.6-flash (zero-shot, val) | 61.4% | Official leaderboard | ||
| 3 | gpt-6.2-astra (zero-shot, val) | 60.5% | Official leaderboard | ||
| 4 | gpt-6.2-sol (zero-shot, val) | 60.3% | Official leaderboard | ||
| 5 | gemini/gemini-3.5-flash-lite (zero-shot, val) | 59.7% | Official leaderboard | ||
| 6 | mistralai/Mistral-Large-3-675B-Instruct-2512 (val) | 58.6% | Official leaderboard | ||
| 7 | openai/gpt-5.6-luna (zero-shot, val) | 57.8% | Official leaderboard | ||
| 8 | gemini/gemini-3-flash-preview#thinking (zero-shot, val) | 57.4% | Official leaderboard | ||
| 9 | NbAiLab/borealis2-31b-preview (val) | 57.3% | Official leaderboard | ||
| 10 | claude-sonnet-4-5-20250929#no-thinking (zero-shot, val) | 57.3% | Official leaderboard | ||
| 11 | gpt-4.1-2025-04-14 (zero-shot, val) | 56.4% | Official leaderboard | ||
| 12 | deepseek/deepseek-flash#no-thinking (val) | 56.2% | Official leaderboard | ||
| 13 | gpt-6.1-sol (zero-shot, val) | 56.1% | Official leaderboard | ||
| 14 | utter-project/EuroLLM-22B-Instruct-2512 | 54.7% | Official leaderboard | ||
| 15 | zai-org/GLM-5.3-Flash (val) | 54.6% | Official leaderboard | ||
| 16 | gpt-6-astra (zero-shot, val) | 54.3% | Official leaderboard | ||
| 17 | gpt-6-sol (zero-shot, val) | 54.2% | Official leaderboard | ||
| 18 | gpt-5-mini-2025-08-07 (zero-shot, val) | 54.1% | Official leaderboard | ||
| 19 | gemini/gemini-3.1-flash-lite-preview (zero-shot, val) | 54% | Official leaderboard | ||
| 20 | mistralai/Mistral-Medium-3.5-128B (val) | 53.7% | Official leaderboard | ||
| 21 | google/gemma-4-31B-it (val) | 53% | Official leaderboard | ||
| 22 | claude-sonnet-4-5-20250929#thinking (zero-shot, val) | 52.9% | Official leaderboard | ||
| 23 | xai/grok-4-1-fast-reasoning (zero-shot, val) | 52.6% | Official leaderboard | ||
| 24 | mistralai/Mistral-Small-4-119B-2603 (val) | 52.2% | Official leaderboard | ||
| 25 | claude-sonnet-4-6 (zero-shot, val) | 51.6% | Official leaderboard | ||
| 26 | openai/gpt-5.4-mini-2026-03-17#none (zero-shot, val) | 51.6% | Official leaderboard | ||
| 27 | openai/gpt-5.4-mini-2026-03-17#low (zero-shot, val) | 51.3% | Official leaderboard | ||
| 28 | Qwen/Qwen3.6-27B-FP8 (val) | 51% | Official leaderboard | ||
| 29 | deepcogito/cogito-v1-preview-qwen-14B | 50.8% | Official leaderboard | ||
| 30 | mistralai/Mistral-Medium-3.5-128B (zero-shot, val) | 50.6% | Official leaderboard | ||
| 31 | openai/gpt-5.4-2026-03-05#none (zero-shot, val) | 50.6% | Official leaderboard | ||
| 32 | NbAiLab/borealis2-26b-a4b-preview (val) | 49.8% | Official leaderboard | ||
| 33 | speakleash/Bielik-11B-v3.1-Instruct (val) | 49.7% | Official leaderboard | ||
| 34 | mistralai/Ministral-3-14B-Base-2512 (val) | 49.6% | Official leaderboard | ||
| 35 | deepcogito/cogito-v2-preview-llama-70B | 49.6% | Official leaderboard | ||
| 36 | mistralai/Ministral-3-14B-Reasoning-2512 (val) | 49.5% | Official leaderboard | ||
| 37 | | 49.4% | Mar 2025 | Official leaderboard | |
| 38 | google/gemma-4-26B-A4B-it (val) | 49.1% | Official leaderboard | ||
| 39 | swiss-ai/Apertus-v1.5-70B (val) | 49% | Official leaderboard | ||
| 40 | Qwen/Qwen3.5-35B-A3B-FP8 (val) | 48.9% | Official leaderboard | ||
| 41 | gpt-6-luna (zero-shot, val) | 48.8% | Official leaderboard | ||
| 42 | NbAiLab/borealis2-26b-a4b-preview (zero-shot, val) | 48.7% | Official leaderboard | ||
| 43 | zai-org/GLM-5.3-Flash (zero-shot, val) | 48.7% | Official leaderboard | ||
| 44 | | 48.5% | Mar 2025 | Official leaderboard | |
| 45 | | 48.2% | Mar 2025 | Official leaderboard | |
| 46 | Qwen/Qwen3-14B#no-thinking | 48% | Official leaderboard | ||
| 47 | Qwen/Qwen3-32B#no-thinking | 48% | Official leaderboard | ||
| 48 | NbAiLab/borealis2-31b-preview (zero-shot, val) | 47.9% | Official leaderboard | ||
| 49 | gpt-5.6-sol (zero-shot, val) | 47.7% | Official leaderboard | ||
| 50 | deepcogito/cogito-v1-preview-qwen-32B | 47.5% | Official leaderboard | ||
| 51 | gemini/gemini-2.5-flash-lite#no-thinking (zero-shot, val) | 47.3% | Official leaderboard | ||
| 52 | amalia-llm/AMALIA-9B-0626-DPO (val) | 46.6% | Official leaderboard | ||
| 53 | deepseek/deepseek-flash#low (zero-shot, val) | 46.4% | Official leaderboard | ||
| 54 | gpt-5.2-2025-12-11 (zero-shot, val) | 46.3% | Official leaderboard | ||
| 55 | | 45.8% | Mar 2025 | Official leaderboard | |
| 56 | claude-haiku-4-5-20251001 (zero-shot, val) | 45.2% | Official leaderboard | ||
| 57 | gemini/gemini-2.5-flash-lite#thinking (zero-shot, val) | 45.1% | Official leaderboard | ||
| 58 | mistralai/Mistral-Small-3.1-24B-Base-2503 | 43.9% | Official leaderboard | ||
| 59 | swiss-ai/Apertus-v1.5-8B (val) | 43.9% | Official leaderboard | ||
| 60 | openai/gpt-oss-120b (val) | 43.7% | Official leaderboard | ||
| 61 | speakleash/Bielik-11B-v3.0-Instruct (val) | 43.1% | Official leaderboard | ||
| 62 | openeurollm/prelude-long (val) | 43.1% | Official leaderboard | ||
| 63 | deepcogito/cogito-v1-preview-llama-8B | 42.6% | Official leaderboard | ||
| 64 | nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 (val) | 42.5% | Official leaderboard | ||
| 65 | nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-Base-BF16 (val) | 42.5% | Official leaderboard | ||
| 66 | latam-gpt/Llama-3.1-70B-LatamGPT-SFT-1.0 (val) | 42.2% | Official leaderboard | ||
| 67 | Qwen/Qwen3-8B#no-thinking | 42.1% | Official leaderboard | ||
| 68 | Qwen/Qwen3.5-9B (val) | 42% | Official leaderboard | ||
| 69 | gpt-5-nano-2025-08-07 (zero-shot, val) | 41.9% | Official leaderboard | ||
| 70 | openai/gpt-5.4-mini-2026-03-17#medium (zero-shot, val) | 41.6% | Official leaderboard | ||
| 71 | openai/gpt-5.4-nano-2026-03-17#medium (zero-shot, val) | 41.6% | Official leaderboard | ||
| 72 | allenai/Llama-3.1-Tulu-3-8B-SFT | 41.3% | Official leaderboard | ||
| 73 | google/gemma-4-12B-it (val) | 41.3% | Official leaderboard | ||
| 74 | mistralai/Mistral-Small-3.2-24B-Instruct-2506 | 41.2% | Official leaderboard | ||
| 75 | openeurollm/prelude-base (val) | 40.9% | Official leaderboard | ||
| 76 | Qwen/Qwen3.8-27B (val) | 40.7% | Official leaderboard | ||
| 77 | gpt-5.6-terra (zero-shot, val) | 40.2% | Official leaderboard | ||
| 78 | speakleash/Bielik-Minitron-7B-v3.0-Instruct (val) | 39.9% | Official leaderboard | ||
| 79 | openai/gpt-5.4-nano-2026-03-17#low (zero-shot, val) | 39.8% | Official leaderboard | ||
| 80 | Qwen/Qwen3-4B#no-thinking | 39.8% | Official leaderboard | ||
| 81 | google/gemma-4-E4B-it (val) | 39.5% | Official leaderboard | ||
| 82 | mistralai/Mistral-Large-3-675B-Instruct-2512 (zero-shot, val) | 39.3% | Official leaderboard | ||
| 83 | speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c (zero-shot) | 39.1% | Official leaderboard | ||
| 84 | nvidia/Nemotron-Cascade-2-30B-A3B (val) | 38.9% | Official leaderboard | ||
| 85 | zai-org/GLM-4.7-Flash (val) | 38.7% | Official leaderboard | ||
| 86 | swiss-ai/Apertus-v1.5-70B (zero-shot, val) | 38.6% | Official leaderboard | ||
| 87 | domyn/Domyn-Small-v1.0 (val) | 38.1% | Official leaderboard | ||
| 88 | mistralai/Ministral-3-8B-Reasoning-2512 | 38% | Official leaderboard | ||
| 89 | | 38% | Jul 2024 | Official leaderboard | |
| 90 | | 37.8% | Mar 2025 | Official leaderboard | |
| 91 | nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8 (val) | 37.6% | Official leaderboard | ||
| 92 | microsoft/mdeberta-v3-base | 37.5% | Official leaderboard | ||
| 93 | meta-llama/Llama-3.1-8B-Instruct | 37.1% | Official leaderboard | ||
| 94 | amalia-llm/AMALIA-9B-0626-SFT (val) | 37.1% | Official leaderboard | ||
| 95 | openai/gpt-5.4-nano-2026-03-17#high (zero-shot, val) | 37% | Official leaderboard | ||
| 96 | mistralai/Mistral-Small-3.1-24B-Instruct-2503 | 37% | Official leaderboard | ||
| 97 | mistralai/Mistral-Small-4-119B-2603 (zero-shot, val) | 36.6% | Official leaderboard | ||
| 98 | gpt-oss-20b | 36.3% | low | Aug 2025 | Official leaderboard |
| 99 | Qwen/Qwen3.5-9B-Base (val) | 36.3% | Official leaderboard | ||
| 100 | mistralai/Ministral-3-14B-Instruct-2512 (val) | 36.3% | Official leaderboard | ||
| 101 | openai/gpt-5.4-mini-2026-03-17#high (zero-shot, val) | 36.1% | Official leaderboard | ||
| 102 | OpenLLM-France/Luciole-23B-Instruct-1.1 (val) | 35.9% | Official leaderboard | ||
| 103 | Qwen/Qwen3.5-35B-A3B-Base (val) | 35.9% | Official leaderboard | ||
| 104 | speakleash/Bielik-11B-v2.3-Instruct | 35.2% | Official leaderboard | ||
| 105 | NbAiLab/borealis2-e4b-preview (val) | 35.2% | Official leaderboard | ||
| 106 | CohereLabs/aya-expanse-8b (val) | 34.7% | Official leaderboard | ||
| 107 | utter-project/EuroLLM-9B-Instruct | 33.9% | Official leaderboard | ||
| 108 | | 33.6% | Aug 2025 | Official leaderboard | |
| 109 | utter-project/EuroLLM-9B-Instruct-2512 (val) | 33.1% | Official leaderboard | ||
| 110 | OpenLLM-France/Luciole-8B-Instruct-1.1 (val) | 32.4% | Official leaderboard | ||
| 111 | google/gemma-4-31B (val) | 32.2% | Official leaderboard | ||
| 112 | Qwen/Qwen3.5-2B (val) | 31.6% | Official leaderboard | ||
| 113 | granite-4.0-micro | 31.5% | Oct 2025 | Official leaderboard | |
| 114 | NbAiLab/borealis2-e2b-preview (val) | 31.1% | Official leaderboard | ||
| 115 | | 31.1% | Mar 2025 | Official leaderboard | |
| 116 | mii-llm/nesso-4B | 31% | Official leaderboard | ||
| 117 | allenai/Olmo-3-1125-32B | 30.7% | Official leaderboard | ||
| 118 | danish-foundation-models/DFM-Mimir-v1.8 (val) | 30.4% | Official leaderboard | ||
| 119 | Aleph-Alpha/Kolibri-1 (val) | 29.8% | Official leaderboard | ||
| 120 | google/gemma-4-E2B-it (val) | 29.8% | Official leaderboard | ||
| 121 | typesafe/jev-1.13.0 (zero-shot) | 29.8% | Official leaderboard | ||
| 122 | mistralai/Ministral-3-3B-Instruct-2512 (val) | 29.4% | Official leaderboard | ||
| 123 | deepcogito/cogito-v1-preview-llama-3B | 29% | Official leaderboard | ||
| 124 | allenai/Olmo-3.1-32B-Instruct-DPO | 28.8% | Official leaderboard | ||
| 125 | allenai/Olmo-3.1-32B-Instruct-SFT | 28.8% | Official leaderboard | ||
| 126 | danish-foundation-models/DFM-Mimir-v1.8 (zero-shot, val) | 28.5% | Official leaderboard | ||
| 127 | | 27.5% | Dec 2025 | Official leaderboard | |
| 128 | ibm-granite/granite-4.0-h-micro | 27.2% | Official leaderboard | ||
| 129 | Qwen/Qwen3.5-4B (val) | 27.1% | Official leaderboard | ||
| 130 | | 26.9% | Dec 2025 | Official leaderboard | |
| 131 | | 26.8% | Jul 2024 | Official leaderboard | |
| 132 | danish-foundation-models/DFM-Mimir-v1.5 (val) | 26.6% | Official leaderboard | ||
| 133 | danish-foundation-models/DFM-Mimir-v1.5 (zero-shot, val) | 25.9% | Official leaderboard | ||
| 134 | BSC-LT/salamandra-7b-instruct | 25.8% | Official leaderboard | ||
| 135 | amalia-llm/AMALIA-9B-0626-DPO (zero-shot, val) | 25.5% | Official leaderboard | ||
| 136 | amalia-llm/AMALIA-9B-0626-SFT (zero-shot, val) | 24.8% | Official leaderboard | ||
| 137 | swiss-ai/Apertus-v1.1-4B (val) | 24.6% | Official leaderboard | ||
| 138 | google/gemma-4-26B-A4B (val) | 23.8% | Official leaderboard | ||
| 139 | Qwen/Qwen3.5-2B-Base (val) | 22.8% | Official leaderboard | ||
| 140 | meta-llama/Llama-3.2-3B-Instruct | 22.6% | Official leaderboard | ||
| 141 | inclusionAI/Ling-mini-2.0 (val) | 22.5% | Official leaderboard | ||
| 142 | OpenLLM-France/Luciole-23B-Base (val) | 22.3% | Official leaderboard | ||
| 143 | google/gemma-4-E4B (val) | 22% | Official leaderboard | ||
| 144 | google/gemma-4-12B (val) | 21.4% | Official leaderboard | ||
| 145 | mistralai/Ministral-3-8B-Instruct-2512 (val) | 21.3% | Official leaderboard | ||
| 146 | Qwen/Qwen3.5-4B-Base (val) | 21.2% | Official leaderboard | ||
| 147 | LGAI-EXAONE/EXAONE-4.0-32B (val) | 20.5% | Official leaderboard | ||
| 148 | Qwen/Qwen3-1.7B#no-thinking | 20.3% | Official leaderboard | ||
| 149 | utter-project/EuroLLM-22B-2512 | 20.1% | Official leaderboard | ||
| 150 | nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16 | 19.9% | Official leaderboard | ||
| 151 | HuggingFaceTB/SmolLM3-3B-Base | 19.9% | Official leaderboard | ||
| 152 | nvidia/NVIDIA-Nemotron-3-Nano-4B-FP8 (val) | 19.8% | Official leaderboard | ||
| 153 | danish-foundation-models/DFM-Mimir (val) | 19.6% | Official leaderboard | ||
| 154 | Qwen/Qwen3.5-0.8B-Base (val) | 19.3% | Official leaderboard | ||
| 155 | openai/gpt-5.4-nano-2026-03-17#none (zero-shot, val) | 19% | Official leaderboard | ||
| 156 | jhu-clsp/mmBERT-small | 18.5% | Official leaderboard | ||
| 157 | jhu-clsp/mmBERT-base | 18.4% | Official leaderboard | ||
| 158 | swiss-ai/Apertus-v1.5-8B (zero-shot, val) | 18.1% | Official leaderboard | ||
| 159 | allenai/Olmo-3-1025-7B | 17.6% | Official leaderboard | ||
| 160 | speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551 (zero-shot) | 17.3% | Official leaderboard | ||
| 161 | danish-foundation-models/DFM-Mimir-v1.8-Mini (val) | 17.1% | Official leaderboard | ||
| 162 | JetBrains/Mellum2-12B-A2.5B-Instruct (val) | 16.7% | Official leaderboard | ||
| 163 | openGPT-X/Teuken-7B-instruct-v0.6 (val) | 16.2% | Official leaderboard | ||
| 164 | HuggingFaceTB/SmolLM3-3B#no-thinking | 15.7% | Official leaderboard | ||
| 165 | mistralai/Ministral-3-8B-Base-2512 (val) | 15.7% | Official leaderboard | ||
| 166 | occiglot/occiglot-7b-eu5-instruct | 15.5% | Official leaderboard | ||
| 167 | JetBrains/Mellum2-12B-A2.5B-Base (val) | 15.2% | Official leaderboard | ||
| 168 | Qwen/Qwen3.5-0.8B (val) | 14.9% | Official leaderboard | ||
| 169 | swiss-ai/Apertus-v1.1-4B-Instruct (val) | 14.9% | Official leaderboard | ||
| 170 | danish-foundation-models/DFM-Mimir-v1.8-Mini (zero-shot, val) | 14.8% | Official leaderboard | ||
| 171 | allenai/Olmo-3-7B-Instruct | 14.3% | Official leaderboard | ||
| 172 | openeurollm/oellm-9b-128k-theta32m-v3 (val) | 13.9% | Official leaderboard | ||
| 173 | allenai/Olmo-3-7B-Instruct-DPO | 13.6% | Official leaderboard | ||
| 174 | danish-foundation-models/DFM-Mimir (zero-shot, val) | 13.6% | Official leaderboard | ||
| 175 | LiquidAI/LFM2.5-8B-A1B (val) | 12.6% | Official leaderboard | ||
| 176 | intfloat/multilingual-e5-base | 12.4% | Official leaderboard | ||
| 177 | gemini/gemini-3-flash-preview#no-thinking (zero-shot, val) | 11.8% | Official leaderboard | ||
| 178 | intfloat/multilingual-e5-large | 11.4% | Official leaderboard | ||
| 179 | allenai/Olmo-3-7B-Instruct-SFT | 11.3% | Official leaderboard | ||
| 180 | EuroBERT/EuroBERT-210m | 11.3% | Official leaderboard | ||
| 181 | LiquidAI/LFM2.5-8B-A1B-Base (val) | 11.3% | Official leaderboard | ||
| 182 | answerdotai/ModernBERT-large | 11.2% | Official leaderboard | ||
| 183 | EuroBERT/EuroBERT-610m | 11.1% | Official leaderboard | ||
| 184 | CohereLabs/aya-23-8B (val) | 9.84% | Official leaderboard | ||
| 185 | Aleph-Alpha/Kolibri-1 (zero-shot, val) | 9.65% | Official leaderboard | ||
| 186 | speakleash/Bielik-1.5B-v3.0-Instruct | 9.52% | Official leaderboard | ||
| 187 | utter-project/EuroLLM-9B | 9.41% | Official leaderboard | ||
| 188 | TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49 | 8.80% | Official leaderboard | ||
| 189 | distilbert/distilbert-base-multilingual-cased | 7.68% | Official leaderboard | ||
| 190 | Geotrend/distilbert-base-25lang-cased | 7.56% | Official leaderboard | ||
| 191 | google/gemma-4-E2B (val) | 7.55% | Official leaderboard | ||
| 192 | | 7.00% | Apr 2025 | Official leaderboard | |
| 193 | answerdotai/ModernBERT-base | 6.56% | Official leaderboard | ||
| 194 | AI-Sweden-Models/bert-large-nordic-pile-1M-steps | 6.47% | Official leaderboard | ||
| 195 | swiss-ai/Apertus-v1.1-1.5B (val) | 6.40% | Official leaderboard | ||
| 196 | AI-Sweden-Models/Llama-3-8B-instruct (val) | 6.02% | Official leaderboard | ||
| 197 | OmniJev/OneJev1.6-0.8B (zero-shot) | 5.99% | Official leaderboard | ||
| 198 | | 5.90% | Mar 2025 | Official leaderboard | |
| 199 | EuropeanParliament/EUBERT | 5.74% | Official leaderboard | ||
| 200 | OpenLLM-France/Luciole-8B-Base (val) | 5.50% | Official leaderboard | ||
| 201 | TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f (zero-shot) | 5.00% | Official leaderboard | ||
| 202 | TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894 | 4.90% | Official leaderboard | ||
| 203 | fresh-electra-small | 4.70% | Official leaderboard | ||
| 204 | fresh-xlm-roberta-base | 4.62% | Official leaderboard | ||
| 205 | flwrlabs/Lizzy-7B (val) | 4.61% | Official leaderboard | ||
| 206 | NbAiLab/borealis2-e4b-preview (zero-shot, val) | 4.37% | Official leaderboard | ||
| 207 | syvai/claes-v1 (zero-shot) | 4.30% | Official leaderboard | ||
| 208 | convaiinnovations/laya (zero-shot) | 4.00% | Official leaderboard | ||
| 209 | occiglot/occiglot-7b-eu5 | 3.88% | Official leaderboard | ||
| 210 | Qwen/Qwen3-0.6B#no-thinking | 3.60% | Official leaderboard | ||
| 211 | LFM2-1.2B | 3.52% | Jul 2025 | Official leaderboard | |
| 212 | tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base | 3.37% | Official leaderboard | ||
| 213 | TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660 (zero-shot) | 3.34% | Official leaderboard | ||
| 214 | JetBrains/Mellum2-12B-A2.5B-Thinking-SFT (val) | 3.09% | Official leaderboard | ||
| 215 | PleIAs/Pleias-350m-Preview | 2.98% | Official leaderboard | ||
| 216 | yulan-team/YuLan-Mini-Instruct | 2.90% | Official leaderboard | ||
| 217 | meta-llama/Llama-3.2-1B-Instruct | 2.78% | Official leaderboard | ||
| 218 | TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6 | 2.62% | Official leaderboard | ||
| 219 | | 2.52% | Mar 2025 | Official leaderboard | |
| 220 | | 2.36% | Aug 2025 | Official leaderboard | |
| 221 | | 2.35% | Sep 2024 | Official leaderboard | |
| 222 | BSC-LT/salamandra-7b | 1.95% | Official leaderboard | ||
| 223 | PleIAs/Pleias-1.2b-Preview | 1.91% | Official leaderboard | ||
| 224 | JetBrains/Mellum2-12B-A2.5B-Thinking (val) | 1.62% | Official leaderboard | ||
| 225 | LiquidAI/LFM2-350M | 1.08% | Official leaderboard | ||
| 226 | | 0.70% | Oct 2024 | Official leaderboard | |
| 227 | | 0.46% | Oct 2024 | Official leaderboard | |
| 228 | HuggingFaceTB/SmolLM2-135M-Instruct | 0.40% | Official leaderboard | ||
| 229 | google-bert/bert-base-multilingual-cased (val) | 0.23% | Official leaderboard | ||
| 230 | LiquidAI/LFM2-700M | 0.19% | Official leaderboard | ||
| 231 | jaredpalmer/kev-0.5b (zero-shot) | 0.14% | Official leaderboard | ||
| 232 | HuggingFaceTB/SmolLM2-360M-Instruct | 0.06% | Official leaderboard | ||
| 233 | AI-Sweden-Models/Llama-3-8B-instruct (zero-shot, val) | 0% | Official leaderboard | ||
| 234 | JetBrains/Mellum2-12B-A2.5B-Base-Pretrain (val) | 0% | Official leaderboard | ||
| 235 | mistralai/Ministral-3-14B-Instruct-2512 (zero-shot, val) | 0% | Official leaderboard | ||
| 236 | mistralai/Ministral-3-3B-Instruct-2512 (zero-shot, val) | 0% | Official leaderboard | ||
| 237 | Muse-research/Muse-3B (val) | 0% | Official leaderboard | ||
| 238 | openGPT-X/Teuken-7B-base-v0.6 (val) | 0% | Official leaderboard | ||
| 239 | PleIAs/Baguettotron@refs/pr/6#no-thinking | 0% | Official leaderboard | ||
| 240 | swiss-ai/Apertus-v1.1-1.5B-Instruct (val) | 0% | Official leaderboard | ||
| 241 | utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview | 0% | Official leaderboard | ||
| 242 | tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct | -0.06% | Official leaderboard | ||
| 243 | swiss-ai/Apertus-v1.1-1.5B (zero-shot) | -0.15% | Official leaderboard | ||
| 244 | NbAiLab/borealis2-e2b-preview (zero-shot, val) | -0.23% | Official leaderboard | ||
| 245 | convaiinnovations/laya-multilingual (zero-shot) | -0.32% | Official leaderboard | ||
| 246 | TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660 | -0.56% | Official leaderboard | ||
| 247 | JetBrains/Mellum2-12B-A2.5B-Instruct-SFT (val) | -0.92% | Official leaderboard | ||
| 248 | PleIAs/Monad | -1.23% | Official leaderboard | ||
| 249 | AIDC-AI/Marco-Mini-Global-Base (val) | -1.34% | Official leaderboard | ||
| 250 | syvai/claes-preview (zero-shot) | -1.76% | Official leaderboard | ||
| 251 | TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6 (zero-shot) | -1.82% | Official leaderboard | ||
| 252 | OpenLLM-France/Luciole-1B-Base (val) | -2.26% | Official leaderboard | ||
| 253 | OpenLLM-France/Luciole-1B-Instruct-1.1 (val) | -3.56% | Official leaderboard | ||
| 254 | TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f | -3.72% | Official leaderboard | ||
| 255 | swiss-ai/Apertus-v1.1-0.5B-Instruct (val) | -3.89% | Official leaderboard | ||
| 256 | fastino/GLiNER2.5-Decide (zero-shot) | -5.92% | Official leaderboard | ||
| 257 | swiss-ai/Apertus-v1.1-0.5B (val) | -6.66% | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 49.4% | Mar 2025 | Official leaderboard | |
| 2 | | 48.5% | Mar 2025 | Official leaderboard | |
| 3 | | 48.2% | Mar 2025 | Official leaderboard | |
| 4 | | 45.8% | Mar 2025 | Official leaderboard | |
| 5 | | 38% | Jul 2024 | Official leaderboard | |
| 6 | | 37.8% | Mar 2025 | Official leaderboard | |
| 7 | gpt-oss-20b | 36.3% | low | Aug 2025 | Official leaderboard |
| 8 | | 33.6% | Aug 2025 | Official leaderboard | |
| 9 | granite-4.0-micro | 31.5% | Oct 2025 | Official leaderboard | |
| 10 | | 31.1% | Mar 2025 | Official leaderboard | |
| 11 | | 27.5% | Dec 2025 | Official leaderboard | |
| 12 | | 26.9% | Dec 2025 | Official leaderboard | |
| 13 | | 26.8% | Jul 2024 | Official leaderboard | |
| 14 | | 7.00% | Apr 2025 | Official leaderboard | |
| 15 | | 5.90% | Mar 2025 | Official leaderboard | |
| 16 | LFM2-1.2B | 3.52% | Jul 2025 | Official leaderboard | |
| 17 | | 2.52% | Mar 2025 | Official leaderboard | |
| 18 | | 2.36% | Aug 2025 | Official leaderboard | |
| 19 | | 2.35% | Sep 2024 | Official leaderboard | |
| 20 | | 0.70% | Oct 2024 | Official leaderboard | |
| 21 | | 0.46% | Oct 2024 | Official leaderboard |
2 models have held number one on the official leaderboard since its release in April 2024
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: Official Dutch CoLA leaderboard
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official Dutch CoLA 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
- 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 11 months ago, in December 2025, for Olmo 3.1 32B Instruct. 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 (257 entries).
- Pass Models scored 21 models we track, plus 236 other entries on its leaderboard.
- Warning Newest score The newest came 11 months ago, in December 2025, for Olmo 3.1 32B Instruct.
- Note Frontier models None of the top 10 has a score. The strongest tested is gpt-oss-20b (#143).
- Pass Headroom The best is 49.4%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on euroeval.com: 24 rows for models we track and 236 other entries at the last read.
- Pass Last read November 21st, 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.