DaLA
Models distinguish acceptable Danish sentences from copies corrupted to resemble errors in real text.
EuroEval compares the labels with human or constructed reference judgments using the Matthews correlation coefficient.
- Made by
- Gianluca Barmina and colleagues
- Released
- December 2025
- Kind
- Reasoning in many languages
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Multiple choice
- Tasks
- All public
- Measures
- Capability
- Subject
- Languages and literature
- Language
- Danish
- Interaction
- Single turn
- Items from
- Generated
- Refresh
- One fixed set
- Run by
- Anyone
- Strongest model tested
-
Qwen3.5-9B #128 on the capability index
- Last model added
- February 2026
- Scores refreshed
- November 20th, 2026
AI-Sweden-Models/roberta-large-1160k tops the leaderboard with 69% Gemma 3 27B PT leads with 61.7% among models released by February 2026
Best score per entry from the benchmark's paper, read on November 20th, 2026. A paper’s table doesn’t add newer models. 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 PT, at 61.7%. This view shows only models we track, each at its best.
The newest model with a score here came out in February 2026, 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
- 69.0 AI-Sweden-Models/roberta-large-1160k
- 65.9 microsoft/mdeberta-v3-base
- 64.3 utter-project/EuroLLM-22B-Instruct-2512
- 61.7
Gemma 3 27B PT
- 60.3 AI-Sweden-Models/ModernBERT-large
- 56.7
Gemma 3 27B
- 54.3 jhu-clsp/mmBERT-base
- 52.3
Gemma 3 12B PT
- 50.9
Gemma 3 12B
- 49.4
Qwen3-32B
- 49.1 danish-foundation-models/munin-qwen3.5-9B
- 45.5
Qwen3.5-9B
- 42.3 gpt-oss-20b
- 41.5 deepcogito/cogito-v2-preview-llama-70B
- 41.3 jhu-clsp/mmBERT-small
- 41.0
Qwen3-14B
- 40.8 deepcogito/cogito-v1-preview-qwen-32B
- 40.7 mistralai/Mistral-Small-3.1-24B-Base-2503
- 38.6
Mistral Small 3.2
- 38.5
Mistral Small 3.1
- 37.8 utter-project/EuroLLM-22B-2512
- 37.4 mistralai/Ministral-3-8B-Reasoning-2512
- 35.2 deepcogito/cogito-v1-preview-qwen-14B
- 33.4
Llama 3.1-70B
- 33.4 BSC-LT/ALIA-40b
- 32.2
Qwen3-8B
- 32.0
Qwen3-4B-Thinking-2507
- 31.2 speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c
- 30.0 utter-project/EuroLLM-9B
- 29.9 typesafe/jev-1.13.0
- 29.6 FacebookAI/xlm-roberta-large
- 29.2
Qwen3-4B
- 28.4 utter-project/EuroLLM-9B-Instruct
- 26.7 deepcogito/cogito-v1-preview-llama-8B
- 25.9 meta-llama/Llama-3.1-8B-Instruct
- 25.1 danish-foundation-models/munin-apertus-8b
- 23.9 mii-llm/nesso-4B
- 21.8 allenai/Llama-3.1-Tulu-3-8B-SFT
- 20.1 speakleash/Bielik-11B-v2.3-Instruct
- 19.0 granite-4.0-micro
- 18.1 allenai/Olmo-3.1-32B-Instruct-DPO
- 17.9
Qwen3-1.7B
- 17.8 allenai/Olmo-3.1-32B-Instruct-SFT
- 17.5 danish-foundation-models/munin-gemma4-e4b
- 17.0 danish-foundation-models/munin-ministral3-8B
- 17.0 speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551
- 16.9 BSC-LT/salamandra-7b-instruct
- 16.5 intfloat/multilingual-e5-large
- 16.1
Olmo 3.1 32B Instruct
- 16.0 BSC-LT/salamandra-7b
- 15.2 allenai/Olmo-3-1125-32B
- 15.1 ibm-granite/granite-4.0-h-micro
- 14.5 nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16
- 13.7 bineric/lynx-instruct-30b
- 13.1
Ministral 3 3B Reasoning
- 12.9 synquid/gemma-3-4b-pt-laerebogen-with-follow-ups-lora
- 9.50 allenai/Olmo-3-7B-Instruct
- 8.84
Llama 3.1-8B
- 8.82 deepcogito/cogito-v1-preview-llama-3B
- 8.39 allenai/Olmo-3-7B-Instruct-DPO
- 8.21 allenai/Olmo-3-1025-7B
- 8.05 meta-llama/Llama-3.2-3B-Instruct
- 8.04 allenai/Olmo-3-7B-Instruct-SFT
- 5.76 HuggingFaceTB/SmolLM3-3B
- 5.76
Gemma 3 1B
- 4.72 speakleash/Bielik-1.5B-v3.0-Instruct
- 4.68 OmniJev/OneJev1.6-0.8B
- 4.49 intfloat/multilingual-e5-base
- 4.44 AI-Sweden-Models/bert-large-nordic-pile-1M-steps
- 4.34
Qwen3-0.6B
- 3.69 convaiinnovations/laya
- 3.24 occiglot/occiglot-7b-eu5-instruct
- 3.04 occiglot/occiglot-7b-eu5
- 2.58
Llama 3.2 1B
- 2.43 answerdotai/ModernBERT-large
- 2.08 answerdotai/ModernBERT-base
- 1.74 distilbert/distilbert-base-multilingual-cased
- 1.71 EuroBERT/EuroBERT-210m
- 1.67 Geotrend/distilbert-base-25lang-cased
- 1.60 PleIAs/Pleias-3b-Preview
- 1.59 LiquidAI/LFM2-350M
- 1.56 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660
- 1.52 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f
- 1.50 TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894
- 1.45 TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49
- 1.43 EuroBERT/EuroBERT-610m
- 1.29 fresh-xlm-roberta-base
- 1.22 yulan-team/YuLan-Mini-Instruct
- 1.16 LFM2-1.2B
- 1.03 fastino/GLiNER2.5-Decide
- 0.87 LiquidAI/LFM2-700M
- 0.69 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6
- 0.68
Gemma 3 1B PT
- 0.49 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview
- 0.41 meta-llama/Llama-3.2-1B-Instruct
- 0.31 EuropeanParliament/EUBERT
- 0.19 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct
- 0.08 syvai/claes-v1
- 0.07 convaiinnovations/laya-multilingual
- 0.0 jaredpalmer/kev-0.5b
- 0.0 syvai/claes-preview
- 0.0 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base
- -0.01
Gemma 3 270M
- -0.02 fresh-electra-small
- -0.13 HuggingFaceTB/SmolLM3-3B-Base
- -0.32 PleIAs/Monad
- -0.50 PleIAs/Pleias-350m-Preview
- -0.85 PleIAs/Pleias-1.2b-Preview
- 1 AI-Sweden-Models/roberta-large-1160k 69%
- 2 microsoft/mdeberta-v3-base 65.9%
- 3 utter-project/EuroLLM-22B-Instruct-2512 64.3%
- 4
Gemma 3 27B PT 61.7%
- 5 AI-Sweden-Models/ModernBERT-large 60.3%
- 6
Gemma 3 27B 56.7%
- 7 jhu-clsp/mmBERT-base 54.3%
- 8
Gemma 3 12B PT 52.3%
- 9
Gemma 3 12B 50.9%
- 10
Qwen3-32B no-thinking 49.4%
- 11 danish-foundation-models/munin-qwen3.5-9B 49.1%
- 12
Qwen3.5-9B no-thinking 45.5%
- 13 gpt-oss-20b medium 42.3%
- 14 deepcogito/cogito-v2-preview-llama-70B 41.5%
- 15 jhu-clsp/mmBERT-small 41.3%
- 16
Qwen3-14B no-thinking 41%
- 17 deepcogito/cogito-v1-preview-qwen-32B 40.8%
- 18 mistralai/Mistral-Small-3.1-24B-Base-2503 40.7%
- 19
Mistral Small 3.2 38.6%
- 20
Mistral Small 3.1 38.5%
- 21 utter-project/EuroLLM-22B-2512 37.8%
- 22 mistralai/Ministral-3-8B-Reasoning-2512 37.4%
- 23 deepcogito/cogito-v1-preview-qwen-14B 35.2%
- 24
Llama 3.1-70B 33.4%
- 25 BSC-LT/ALIA-40b 33.4%
- 26
Qwen3-8B no-thinking 32.2%
- 27
Qwen3-4B-Thinking-2507 32%
- 28 speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c zero-shot 31.2%
- 29 utter-project/EuroLLM-9B 30%
- 30 typesafe/jev-1.13.0 zero-shot 29.9%
- 31 FacebookAI/xlm-roberta-large 29.6%
- 32
Qwen3-4B no-thinking 29.2%
- 33 utter-project/EuroLLM-9B-Instruct 28.4%
- 34 deepcogito/cogito-v1-preview-llama-8B 26.7%
- 35 meta-llama/Llama-3.1-8B-Instruct 25.9%
- 36 danish-foundation-models/munin-apertus-8b 25.1%
- 37 mii-llm/nesso-4B 23.9%
- 38 allenai/Llama-3.1-Tulu-3-8B-SFT 21.8%
- 39 speakleash/Bielik-11B-v2.3-Instruct 20.1%
- 40 granite-4.0-micro 19%
- 41 allenai/Olmo-3.1-32B-Instruct-DPO 18.1%
- 42
Qwen3-1.7B no-thinking 17.9%
- 43 allenai/Olmo-3.1-32B-Instruct-SFT 17.8%
- 44 danish-foundation-models/munin-gemma4-e4b 17.5%
- 45 danish-foundation-models/munin-ministral3-8B 17%
- 46 speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551 zero-shot 17%
- 47 BSC-LT/salamandra-7b-instruct 16.9%
- 48 intfloat/multilingual-e5-large 16.5%
- 49
Olmo 3.1 32B Instruct 16.1%
- 50 BSC-LT/salamandra-7b 16%
- 51 allenai/Olmo-3-1125-32B 15.2%
- 52 ibm-granite/granite-4.0-h-micro 15.1%
- 53 nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16 14.5%
- 54 bineric/lynx-instruct-30b no-thinking 13.7%
- 55
Ministral 3 3B Reasoning 13.1%
- 56 synquid/gemma-3-4b-pt-laerebogen-with-follow-ups-lora 12.9%
- 57 allenai/Olmo-3-7B-Instruct 9.50%
- 58
Llama 3.1-8B 8.84%
- 59 deepcogito/cogito-v1-preview-llama-3B 8.82%
- 60 allenai/Olmo-3-7B-Instruct-DPO 8.39%
- 61 allenai/Olmo-3-1025-7B 8.21%
- 62 meta-llama/Llama-3.2-3B-Instruct 8.05%
- 63 allenai/Olmo-3-7B-Instruct-SFT 8.04%
- 64 HuggingFaceTB/SmolLM3-3B no-thinking 5.76%
- 65
Gemma 3 1B 5.76%
- 66 speakleash/Bielik-1.5B-v3.0-Instruct 4.72%
- 67 OmniJev/OneJev1.6-0.8B zero-shot 4.68%
- 68 intfloat/multilingual-e5-base 4.49%
- 69 AI-Sweden-Models/bert-large-nordic-pile-1M-steps 4.44%
- 70
Qwen3-0.6B 4.34%
- 71 convaiinnovations/laya zero-shot 3.69%
- 72 occiglot/occiglot-7b-eu5-instruct 3.24%
- 73 occiglot/occiglot-7b-eu5 3.04%
- 74
Llama 3.2 1B 2.58%
- 75 answerdotai/ModernBERT-large 2.43%
- 76 answerdotai/ModernBERT-base 2.08%
- 77 distilbert/distilbert-base-multilingual-cased 1.74%
- 78 EuroBERT/EuroBERT-210m 1.71%
- 79 Geotrend/distilbert-base-25lang-cased 1.67%
- 80 PleIAs/Pleias-3b-Preview 1.60%
- 81 LiquidAI/LFM2-350M 1.59%
- 82 TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660 1.56%
- 83 TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f 1.52%
- 84 TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894 1.50%
- 85 TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49 1.45%
- 86 EuroBERT/EuroBERT-610m 1.43%
- 87 fresh-xlm-roberta-base 1.29%
- 88 yulan-team/YuLan-Mini-Instruct 1.22%
- 89 LFM2-1.2B 1.16%
- 90 fastino/GLiNER2.5-Decide zero-shot 1.03%
- 91 LiquidAI/LFM2-700M 0.87%
- 92 TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6 0.69%
- 93
Gemma 3 1B PT 0.68%
- 94 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview 0.49%
- 95 meta-llama/Llama-3.2-1B-Instruct 0.41%
- 96 EuropeanParliament/EUBERT 0.31%
- 97 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct 0.19%
- 98 syvai/claes-v1 zero-shot 0.08%
- 99 convaiinnovations/laya-multilingual zero-shot 0.07%
- 100 jaredpalmer/kev-0.5b zero-shot 0%
- 101 syvai/claes-preview zero-shot 0%
- 102 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base 0%
- 103
Gemma 3 270M -0.01%
- 104 fresh-electra-small -0.02%
- 105 HuggingFaceTB/SmolLM3-3B-Base -0.13%
- 106 PleIAs/Monad -0.32%
- 107 PleIAs/Pleias-350m-Preview -0.50%
- 108 PleIAs/Pleias-1.2b-Preview -0.85%
Source: The DaLA paper
- 61.7
Gemma 3 27B PT
- 56.7
Gemma 3 27B
- 52.3
Gemma 3 12B PT
- 50.9
Gemma 3 12B
- 49.4
Qwen3-32B
- 45.5
Qwen3.5-9B
- 42.3 gpt-oss-20b
- 41.0
Qwen3-14B
- 38.6
Mistral Small 3.2
- 38.5
Mistral Small 3.1
- 33.4
Llama 3.1-70B
- 32.2
Qwen3-8B
- 32.0
Qwen3-4B-Thinking-2507
- 29.2
Qwen3-4B
- 19.0 granite-4.0-micro
- 17.9
Qwen3-1.7B
- 16.1
Olmo 3.1 32B Instruct
- 13.1
Ministral 3 3B Reasoning
- 8.84
Llama 3.1-8B
- 5.76
Gemma 3 1B
- 4.34
Qwen3-0.6B
- 2.58
Llama 3.2 1B
- 1.16 LFM2-1.2B
- 0.68
Gemma 3 1B PT
- -0.01
Gemma 3 270M
- 1
Gemma 3 27B PT 61.7%
- 2
Gemma 3 27B 56.7%
- 3
Gemma 3 12B PT 52.3%
- 4
Gemma 3 12B 50.9%
- 5
Qwen3-32B no-thinking 49.4%
- 6
Qwen3.5-9B no-thinking 45.5%
- 7 gpt-oss-20b medium 42.3%
- 8
Qwen3-14B no-thinking 41%
- 9
Mistral Small 3.2 38.6%
- 10
Mistral Small 3.1 38.5%
- 11
Llama 3.1-70B 33.4%
- 12
Qwen3-8B no-thinking 32.2%
- 13
Qwen3-4B-Thinking-2507 32%
- 14
Qwen3-4B no-thinking 29.2%
- 15 granite-4.0-micro 19%
- 16
Qwen3-1.7B no-thinking 17.9%
- 17
Olmo 3.1 32B Instruct 16.1%
- 18
Ministral 3 3B Reasoning 13.1%
- 19
Llama 3.1-8B 8.84%
- 20
Gemma 3 1B 5.76%
- 21
Qwen3-0.6B 4.34%
- 22
Llama 3.2 1B 2.58%
- 23 LFM2-1.2B 1.16%
- 24
Gemma 3 1B PT 0.68%
- 25
Gemma 3 270M -0.01%
Source: The DaLA paper
| # | Entry | Score ± | Setting | Date | Link | Source |
|---|---|---|---|---|---|---|
| 1 | AI-Sweden-Models/roberta-large-1160k open weights | 69% ±1.6 | hf.co | Official leaderboard | ||
| 2 | microsoft/mdeberta-v3-base open weights | 65.9% ±1.7 | hf.co | Official leaderboard | ||
| 3 | utter-project/EuroLLM-22B-Instruct-2512 open weights | 64.3% ±1.1 | hf.co | Official leaderboard | ||
| 4 | | 61.7% ±0.7 | Mar 2025 | hf.co | Official leaderboard | |
| 5 | AI-Sweden-Models/ModernBERT-large open weights | 60.3% ±4.1 | hf.co | Official leaderboard | ||
| 6 | | 56.7% ±0.6 | Mar 2025 | hf.co | Official leaderboard | |
| 7 | jhu-clsp/mmBERT-base open weights | 54.3% ±3.3 | hf.co | Official leaderboard | ||
| 8 | | 52.3% ±1.2 | Mar 2025 | hf.co | Official leaderboard | |
| 9 | | 50.9% ±0.9 | Mar 2025 | hf.co | Official leaderboard | |
| 10 | | 49.4% ±0.6 | no-thinking | Apr 2025 | hf.co | Official leaderboard |
| 11 | danish-foundation-models/munin-qwen3.5-9B open weights | 49.1% ±0.8 | hf.co | Official leaderboard | ||
| 12 | | 45.5% ±1.1 | no-thinking | Feb 2026 | hf.co | Official leaderboard |
| 13 | gpt-oss-20b | 42.3% ±0.7 | medium | Aug 2025 | hf.co | Official leaderboard |
| 14 | deepcogito/cogito-v2-preview-llama-70B open weights | 41.5% ±0.7 | hf.co | Official leaderboard | ||
| 15 | jhu-clsp/mmBERT-small open weights | 41.3% ±5 | hf.co | Official leaderboard | ||
| 16 | | 41% ±0.9 | no-thinking | Apr 2025 | hf.co | Official leaderboard |
| 17 | deepcogito/cogito-v1-preview-qwen-32B open weights | 40.8% ±0.5 | hf.co | Official leaderboard | ||
| 18 | mistralai/Mistral-Small-3.1-24B-Base-2503 open weights | 40.7% ±0.9 | hf.co | Official leaderboard | ||
| 19 | | 38.6% ±1.7 | Jun 2025 | hf.co | Official leaderboard | |
| 20 | | 38.5% ±1 | Mar 2025 | hf.co | Official leaderboard | |
| 21 | utter-project/EuroLLM-22B-2512 open weights | 37.8% ±3.9 | hf.co | Official leaderboard | ||
| 22 | mistralai/Ministral-3-8B-Reasoning-2512 open weights | 37.4% ±0.9 | hf.co | Official leaderboard | ||
| 23 | deepcogito/cogito-v1-preview-qwen-14B open weights | 35.2% ±0.8 | hf.co | Official leaderboard | ||
| 24 | | 33.4% ±2.4 | Jul 2024 | hf.co | Official leaderboard | |
| 25 | BSC-LT/ALIA-40b open weights | 33.4% ±1.5 | hf.co | Official leaderboard | ||
| 26 | | 32.2% ±0.8 | no-thinking | Apr 2025 | hf.co | Official leaderboard |
| 27 | | 32% ±1.2 | Aug 2025 | hf.co | Official leaderboard | |
| 28 | speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c open weights | 31.2% ±0.5 | zero-shot | hf.co | Official leaderboard | |
| 29 | utter-project/EuroLLM-9B open weights | 30% ±1 | hf.co | Official leaderboard | ||
| 30 | typesafe/jev-1.13.0 closed weights | 29.9% ±0.6 | zero-shot | docs.typesafe.ai | Official leaderboard | |
| 31 | FacebookAI/xlm-roberta-large open weights | 29.6% ±9.6 | hf.co | Official leaderboard | ||
| 32 | | 29.2% ±0.8 | no-thinking | Apr 2025 | hf.co | Official leaderboard |
| 33 | utter-project/EuroLLM-9B-Instruct open weights | 28.4% ±1.4 | hf.co | Official leaderboard | ||
| 34 | deepcogito/cogito-v1-preview-llama-8B open weights | 26.7% ±0.7 | hf.co | Official leaderboard | ||
| 35 | meta-llama/Llama-3.1-8B-Instruct open weights | 25.9% ±0.9 | hf.co | Official leaderboard | ||
| 36 | danish-foundation-models/munin-apertus-8b open weights | 25.1% ±2.1 | hf.co | Official leaderboard | ||
| 37 | mii-llm/nesso-4B open weights | 23.9% ±1.1 | hf.co | Official leaderboard | ||
| 38 | allenai/Llama-3.1-Tulu-3-8B-SFT open weights | 21.8% ±1 | hf.co | Official leaderboard | ||
| 39 | speakleash/Bielik-11B-v2.3-Instruct open weights | 20.1% ±1.2 | hf.co | Official leaderboard | ||
| 40 | granite-4.0-micro | 19% ±1 | Oct 2025 | hf.co | Official leaderboard | |
| 41 | allenai/Olmo-3.1-32B-Instruct-DPO open weights | 18.1% ±0.6 | hf.co | Official leaderboard | ||
| 42 | | 17.9% ±0.7 | no-thinking | Apr 2025 | hf.co | Official leaderboard |
| 43 | allenai/Olmo-3.1-32B-Instruct-SFT open weights | 17.8% ±0.7 | hf.co | Official leaderboard | ||
| 44 | danish-foundation-models/munin-gemma4-e4b open weights | 17.5% ±1.7 | hf.co | Official leaderboard | ||
| 45 | danish-foundation-models/munin-ministral3-8B open weights | 17% ±2.3 | hf.co | Official leaderboard | ||
| 46 | speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551 open weights | 17% ±0.5 | zero-shot | hf.co | Official leaderboard | |
| 47 | BSC-LT/salamandra-7b-instruct open weights | 16.9% ±0.9 | hf.co | Official leaderboard | ||
| 48 | intfloat/multilingual-e5-large open weights | 16.5% ±7 | hf.co | Official leaderboard | ||
| 49 | | 16.1% ±0.7 | Dec 2025 | hf.co | Official leaderboard | |
| 50 | BSC-LT/salamandra-7b open weights | 16% ±2 | hf.co | Official leaderboard | ||
| 51 | allenai/Olmo-3-1125-32B open weights | 15.2% ±0.9 | hf.co | Official leaderboard | ||
| 52 | ibm-granite/granite-4.0-h-micro open weights | 15.1% ±0.6 | hf.co | Official leaderboard | ||
| 53 | nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16 open weights | 14.5% ±0.9 | hf.co | Official leaderboard | ||
| 54 | bineric/lynx-instruct-30b open weights | 13.7% ±1 | no-thinking | hf.co | Official leaderboard | |
| 55 | | 13.1% ±1.1 | Dec 2025 | hf.co | Official leaderboard | |
| 56 | synquid/gemma-3-4b-pt-laerebogen-with-follow-ups-lora open weights | 12.9% ±1.1 | hf.co | Official leaderboard | ||
| 57 | allenai/Olmo-3-7B-Instruct open weights | 9.50% ±1.1 | hf.co | Official leaderboard | ||
| 58 | | 8.84% ±1 | Jul 2024 | hf.co | Official leaderboard | |
| 59 | deepcogito/cogito-v1-preview-llama-3B open weights | 8.82% ±0.9 | hf.co | Official leaderboard | ||
| 60 | allenai/Olmo-3-7B-Instruct-DPO open weights | 8.39% ±1 | hf.co | Official leaderboard | ||
| 61 | allenai/Olmo-3-1025-7B open weights | 8.21% ±0.8 | hf.co | Official leaderboard | ||
| 62 | meta-llama/Llama-3.2-3B-Instruct open weights | 8.05% ±0.7 | hf.co | Official leaderboard | ||
| 63 | allenai/Olmo-3-7B-Instruct-SFT open weights | 8.04% ±0.9 | hf.co | Official leaderboard | ||
| 64 | HuggingFaceTB/SmolLM3-3B open weights | 5.76% ±1.1 | no-thinking | hf.co | Official leaderboard | |
| 65 | | 5.76% ±0.6 | Mar 2025 | hf.co | Official leaderboard | |
| 66 | speakleash/Bielik-1.5B-v3.0-Instruct open weights | 4.72% ±1 | hf.co | Official leaderboard | ||
| 67 | OmniJev/OneJev1.6-0.8B open weights | 4.68% ±0.6 | zero-shot | hf.co | Official leaderboard | |
| 68 | intfloat/multilingual-e5-base open weights | 4.49% ±1.5 | hf.co | Official leaderboard | ||
| 69 | AI-Sweden-Models/bert-large-nordic-pile-1M-steps open weights | 4.44% ±2.2 | hf.co | Official leaderboard | ||
| 70 | | 4.34% ±0.7 | Apr 2025 | hf.co | Official leaderboard | |
| 71 | convaiinnovations/laya open weights | 3.69% ±0.6 | zero-shot | hf.co | Official leaderboard | |
| 72 | occiglot/occiglot-7b-eu5-instruct open weights | 3.24% ±0.9 | hf.co | Official leaderboard | ||
| 73 | occiglot/occiglot-7b-eu5 open weights | 3.04% ±1.3 | hf.co | Official leaderboard | ||
| 74 | | 2.58% ±1 | Sep 2024 | hf.co | Official leaderboard | |
| 75 | answerdotai/ModernBERT-large open weights | 2.43% ±1.5 | hf.co | Official leaderboard | ||
| 76 | answerdotai/ModernBERT-base open weights | 2.08% ±1.1 | hf.co | Official leaderboard | ||
| 77 | distilbert/distilbert-base-multilingual-cased open weights | 1.74% ±0.9 | hf.co | Official leaderboard | ||
| 78 | EuroBERT/EuroBERT-210m open weights | 1.71% ±1 | hf.co | Official leaderboard | ||
| 79 | Geotrend/distilbert-base-25lang-cased open weights | 1.67% ±0.9 | hf.co | Official leaderboard | ||
| 80 | PleIAs/Pleias-3b-Preview open weights | 1.60% ±0.8 | hf.co | Official leaderboard | ||
| 81 | LiquidAI/LFM2-350M open weights | 1.59% ±0.5 | hf.co | Official leaderboard | ||
| 82 | TrustLLMeu/trustllm-v2-8b-sft@f442f651172c1a97143738fc30c3a6342d1e8660 open weights | 1.56% ±0.6 | hf.co | Official leaderboard | ||
| 83 | TrustLLMeu/trustllm-v2-8b-sft-8k@2aadff8653393e41abf583222164e279461a619f open weights | 1.52% ±0.7 | hf.co | Official leaderboard | ||
| 84 | TrustLLMeu/trustllm-v2-8b@d8cc924fa550fbf9474f5251e8e2f92d1c1bf894 open weights | 1.50% ±0.8 | hf.co | Official leaderboard | ||
| 85 | TrustLLMeu/trustllm-v2-8b-midtrain@3c7af30bc43bb3c8607e0fab6e7f2a23dd64af49 open weights | 1.45% ±0.8 | hf.co | Official leaderboard | ||
| 86 | EuroBERT/EuroBERT-610m open weights | 1.43% ±0.9 | hf.co | Official leaderboard | ||
| 87 | fresh-xlm-roberta-base open weights | 1.29% ±1 | Official leaderboard | |||
| 88 | yulan-team/YuLan-Mini-Instruct open weights | 1.22% ±0.9 | hf.co | Official leaderboard | ||
| 89 | LFM2-1.2B | 1.16% ±0.7 | Jul 2025 | hf.co | Official leaderboard | |
| 90 | fastino/GLiNER2.5-Decide open weights | 1.03% ±0.8 | zero-shot | hf.co | Official leaderboard | |
| 91 | LiquidAI/LFM2-700M open weights | 0.87% ±1.1 | hf.co | Official leaderboard | ||
| 92 | TrustLLMeu/trustllm-v2-8b-sft-mixb2-exp@f33c260efa853437a55d780ce5b910944b1395f6 open weights | 0.69% ±0.7 | hf.co | Official leaderboard | ||
| 93 | | 0.68% ±0.6 | Mar 2025 | hf.co | Official leaderboard | |
| 94 | utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview open weights | 0.49% ±0.8 | hf.co | Official leaderboard | ||
| 95 | meta-llama/Llama-3.2-1B-Instruct open weights | 0.41% ±0.7 | hf.co | Official leaderboard | ||
| 96 | EuropeanParliament/EUBERT open weights | 0.31% ±0.9 | hf.co | Official leaderboard | ||
| 97 | tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct open weights | 0.19% ±1.2 | hf.co | Official leaderboard | ||
| 98 | syvai/claes-v1 open weights | 0.08% ±0.6 | zero-shot | hf.co | Official leaderboard | |
| 99 | convaiinnovations/laya-multilingual open weights | 0.07% ±0.3 | zero-shot | hf.co | Official leaderboard | |
| 100 | jaredpalmer/kev-0.5b open weights | 0% | zero-shot | hf.co | Official leaderboard | |
| 101 | syvai/claes-preview open weights | 0% | zero-shot | hf.co | Official leaderboard | |
| 102 | tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base open weights | 0% | hf.co | Official leaderboard | ||
| 103 | | -0.01% ±0.4 | Aug 2025 | hf.co | Official leaderboard | |
| 104 | fresh-electra-small open weights | -0.02% ±0.5 | Official leaderboard | |||
| 105 | HuggingFaceTB/SmolLM3-3B-Base open weights | -0.13% ±0.1 | hf.co | Official leaderboard | ||
| 106 | PleIAs/Monad open weights | -0.32% ±0.4 | hf.co | Official leaderboard | ||
| 107 | PleIAs/Pleias-350m-Preview open weights | -0.50% ±0.5 | hf.co | Official leaderboard | ||
| 108 | PleIAs/Pleias-1.2b-Preview open weights | -0.85% ±0.9 | hf.co | Official leaderboard |
| # | Model | Score ± | Setting | Date | Link | Source |
|---|---|---|---|---|---|---|
| 1 | | 61.7% ±0.7 | Mar 2025 | hf.co | Official leaderboard | |
| 2 | | 56.7% ±0.6 | Mar 2025 | hf.co | Official leaderboard | |
| 3 | | 52.3% ±1.2 | Mar 2025 | hf.co | Official leaderboard | |
| 4 | | 50.9% ±0.9 | Mar 2025 | hf.co | Official leaderboard | |
| 5 | | 49.4% ±0.6 | no-thinking | Apr 2025 | hf.co | Official leaderboard |
| 6 | | 45.5% ±1.1 | no-thinking | Feb 2026 | hf.co | Official leaderboard |
| 7 | gpt-oss-20b | 42.3% ±0.7 | medium | Aug 2025 | hf.co | Official leaderboard |
| 8 | | 41% ±0.9 | no-thinking | Apr 2025 | hf.co | Official leaderboard |
| 9 | | 38.6% ±1.7 | Jun 2025 | hf.co | Official leaderboard | |
| 10 | | 38.5% ±1 | Mar 2025 | hf.co | Official leaderboard | |
| 11 | | 33.4% ±2.4 | Jul 2024 | hf.co | Official leaderboard | |
| 12 | | 32.2% ±0.8 | no-thinking | Apr 2025 | hf.co | Official leaderboard |
| 13 | | 32% ±1.2 | Aug 2025 | hf.co | Official leaderboard | |
| 14 | | 29.2% ±0.8 | no-thinking | Apr 2025 | hf.co | Official leaderboard |
| 15 | granite-4.0-micro | 19% ±1 | Oct 2025 | hf.co | Official leaderboard | |
| 16 | | 17.9% ±0.7 | no-thinking | Apr 2025 | hf.co | Official leaderboard |
| 17 | | 16.1% ±0.7 | Dec 2025 | hf.co | Official leaderboard | |
| 18 | | 13.1% ±1.1 | Dec 2025 | hf.co | Official leaderboard | |
| 19 | | 8.84% ±1 | Jul 2024 | hf.co | Official leaderboard | |
| 20 | | 5.76% ±0.6 | Mar 2025 | hf.co | Official leaderboard | |
| 21 | | 4.34% ±0.7 | Apr 2025 | hf.co | Official leaderboard | |
| 22 | | 2.58% ±1 | Sep 2024 | hf.co | Official leaderboard | |
| 23 | LFM2-1.2B | 1.16% ±0.7 | Jul 2025 | hf.co | Official leaderboard | |
| 24 | | 0.68% ±0.6 | Mar 2025 | hf.co | Official leaderboard | |
| 25 | | -0.01% ±0.4 | Aug 2025 | hf.co | Official leaderboard |
Number one over time
Lab claims count from the day the card came out, independent runs from the model’s release.
Source: The DaLA paper
Scores over time
All the scores we have, one dot each, so a model tested at several settings appears more than once.
Source: Official DaLA 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 9 months ago, in February 2026, for Qwen3.5-9B. 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 benchmark’s paper (108 entries).
- Pass Models scored 25 models we track, plus 83 other entries on its leaderboard.
- Warning Newest score The newest came 9 months ago, in February 2026, for Qwen3.5-9B.
- Note Frontier models None of the top 10 has a score. The strongest tested is Qwen3.5-9B (#128).
- Pass Headroom The best is 61.7%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on euroeval.com: 29 rows for models we track and 83 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.