GuiaCat
Models classify the sentiment of restaurant reviews written in Catalan.
EuroEval scores sentiment labels with the Matthews correlation coefficient.
- Made by
- Projecte Aina
- Released
- October 2022
- Kind
- Reading in many languages
- Level
- Everyday
- Scored
- Checked automatically
- Answers
- Multiple choice
- Tasks
- All public
- Measures
- Capability
- Language
- Catalan
- 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
- January 2026
- Scores refreshed
- November 21st, 2026
openeurollm/oellm-9b-128k-theta32m-v3 (val) tops the leaderboard with 78.3% Apertus 70B leads with 69.5% among models released by January 2026
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 Apertus 70B, at 69.5%. This view shows only models we track, each at its best.
The newest model with a score here came out in January 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
- 78.3 openeurollm/oellm-9b-128k-theta32m-v3 (val)
- 77.2 Qwen/Qwen3.5-9B (val)
- 76.9 Qwen/Qwen3.5-27B (val)
- 76.1 openai/gpt-5.4-mini-2026-03-17#high (zero-shot, val)
- 76.0 CohereLabs/tiny-aya-base (val)
- 76.0 zai-org/GLM-4.7-Flash (val)
- 75.9 gpt-5.6-sol (zero-shot, val)
- 75.2 CohereLabs/tiny-aya-global (val)
- 75.1 swiss-ai/Apertus-v1.1-4B (val)
- 74.4 CohereLabs/tiny-aya-water (val)
- 74.1 gemini/gemini-3.6-flash (zero-shot, val)
- 74.0 Qwen/Qwen3.6-27B-FP8 (val)
- 74.0 gemini/gemini-3-flash-preview#thinking (zero-shot, val)
- 74.0 gpt-5.6-terra (zero-shot, val)
- 73.8 gpt-5-mini-2025-08-07 (zero-shot, val)
- 73.8 openai/gpt-5.4-mini-2026-03-17#medium (zero-shot, val)
- 73.7 AI-Sweden-Models/Llama-3-8B-instruct (val)
- 73.7 gemini/gemini-3.7-flash (zero-shot, val)
- 73.6 speakleash/Bielik-11B-v2.3-Instruct (val)
- 73.4 gpt-6.2-sol (zero-shot, val)
- 73.1 Qwen/Qwen3.5-27B-FP8 (val)
- 72.8 Qwen/Qwen3.5-35B-A3B-FP8 (val)
- 72.7 Qwen/Qwen3.8-27B (val)
- 72.6 AI-Sweden-Models/Llama-3-8B-instruct (zero-shot, val)
- 72.5 Qwen/Qwen3.5-4B (val)
- 72.5 GadflyII/GLM-4.7-Flash-NVFP4 (val)
- 72.4 gpt-6-sol (zero-shot, val)
- 72.4 mistralai/Ministral-3-14B-Instruct-2512 (zero-shot, val)
- 72.4 openai/gpt-5.4-mini-2026-03-17#low (zero-shot, val)
- 72.4 google/gemma-4-31B-it (val)
- 72.4 gpt-4.1-2025-04-14 (zero-shot, val)
- 72.2 mistralai/Ministral-3-14B-Base-2512 (val)
- 72.2 claude-sonnet-4-6 (zero-shot, val)
- 72.0 gpt-6-astra (zero-shot, val)
- 72.0 OpenLLM-France/Lucie-7B-Instruct-v1.1 (val)
- 71.9 mistralai/Ministral-3-8B-Base-2512 (val)
- 71.8 mistralai/Ministral-3-14B-Reasoning-2512 (val)
- 71.8 swiss-ai/Apertus-8B-2509
- 71.7 gpt-6.1-sol (zero-shot, val)
- 71.5 google/gemma-4-26B-A4B-it (val)
- 71.5 gpt-5-2025-08-07 (zero-shot, val)
- 71.3 gpt-6.2-astra (zero-shot, val)
- 71.1 swiss-ai/Apertus-v1.1-4B-Instruct (val)
- 71.0 mistralai/Ministral-3-8B-Instruct-2512 (val)
- 71.0 swiss-ai/Apertus-v1.5-70B (zero-shot, val)
- 70.7 latam-gpt/Llama-3.1-70B-LatamGPT-SFT-1.0 (val)
- 70.7 openai/gpt-5.4-nano-2026-03-17#low (zero-shot, val)
- 70.7 openeurollm/prelude-base (val)
- 70.5 mistralai/Mistral-Small-4-119B-2603 (zero-shot, val)
- 70.3 gpt-6-luna (zero-shot, val)
- 70.3 mistralai/Mistral-Medium-3.5-128B (zero-shot, val)
- 70.2 CohereLabs/aya-23-8B (val)
- 69.7 deepseek/deepseek-flash#low (zero-shot, val)
- 69.5 zai-org/GLM-5.3-Flash (val)
- 69.5
Apertus 70B
- 69.5
Llama 3.1-8B
- 69.4 Qwen/Qwen3.5-2B (val)
- 69.4 domyn/Domyn-Small-v1.0 (val)
- 69.3 openai/gpt-5.4-nano-2026-03-17#medium (zero-shot, val)
- 69.0 gemini/gemini-3.5-flash-lite (zero-shot, val)
- 68.9 mistralai/Mistral-Small-4-119B-2603 (val)
- 68.7 gemini/gemini-2.5-pro (zero-shot, val)
- 68.6 gemini/gemini-2.5-flash-lite#thinking (zero-shot, val)
- 68.6 openai/gpt-5.4-mini-2026-03-17#none (zero-shot, val)
- 68.5 gemini/gemini-2.5-flash#thinking (zero-shot, val)
- 68.5 OpenLLM-France/Luciole-8B-Base (val)
- 68.5 NbAiLab/borealis2-31b-preview (zero-shot, val)
- 68.5 claude-haiku-4-5-20251001 (zero-shot, val)
- 68.4 openai/gpt-5.4-2026-03-05#none (zero-shot, val)
- 68.3 gpt-5-nano-2025-08-07 (zero-shot, val)
- 68.3 utter-project/EuroLLM-22B-2512
- 68.2 Unbabel/Tower-Plus-72B (val)
- 68.1 claude-sonnet-4-5-20250929#no-thinking (zero-shot, val)
- 68.1 OpenLLM-France/Luciole-8B-Instruct-1.1 (val)
- 68.1 openai/gpt-5.4-nano-2026-03-17#none (zero-shot, val)
- 68.0 openai/gpt-5.6-luna (zero-shot, val)
- 67.8 xai/grok-4-1-fast-reasoning (zero-shot, val)
- 67.7 gemini/gemini-3.1-flash-lite-preview (zero-shot, val)
- 67.5
Gemma 3 12B PT
- 67.4 openeurollm/prelude-long (val)
- 67.3 NbAiLab/borealis2-31b-preview (val)
- 67.3 nvidia/Nemotron-Cascade-2-30B-A3B (val)
- 67.3 gemini/gemini-3-pro-preview (zero-shot, val)
- 67.2 OpenLLM-France/Luciole-23B-Instruct-1.1 (val)
- 67.2
Llama 3.1-70B
- 67.2 gpt-5.2-2025-12-11 (zero-shot, val)
- 67.1 swiss-ai/Apertus-v1.5-70B (val)
- 67.1 mistralai/Mistral-Medium-3.5-128B (val)
- 67.1 utter-project/EuroLLM-22B-Instruct-2512
- 67.0 openai/gpt-5.4-nano-2026-03-17#high (zero-shot, val)
- 66.9 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 (val)
- 66.7 google/gemma-4-12B (val)
- 66.7 mii-llm/nesso-4B
- 66.5 zai-org/GLM-5.3-Flash (zero-shot, val)
- 66.5 TrustLLMeu/baseline-7-8b_2-3t-tokens_llama
- 66.5 CohereLabs/aya-expanse-8b (val)
- 66.2 mistralai/Mistral-Large-3-675B-Instruct-2512 (zero-shot, val)
- 66.2 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8 (val)
- 66.0 gemini/gemini-2.5-flash#no-thinking (zero-shot, val)
- 66.0 gemini/gemini-3-flash-preview#no-thinking (zero-shot, val)
- 66.0 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-Base-BF16 (val)
- 66.0 danish-foundation-models/DFM-Mimir-v1.8-Mini (zero-shot, val)
- 65.9 OpenLLM-France/Lucie-7B-Instruct-human-data (val)
- 65.8 intfloat/multilingual-e5-large
- 65.5 openai/gpt-oss-120b (val)
- 65.5 mistralai/Ministral-3-3B-Base-2512
- 65.5 danish-foundation-models/DFM-Mimir-v1.8 (zero-shot, val)
- 65.5 utter-project/EuroLLM-22B-Preview
- 65.3 danish-foundation-models/DFM-Mimir (zero-shot, val)
- 65.3 speakleash/Bielik-11B-v2.6-Instruct
- 65.1 mistralai/Ministral-3-14B-Instruct-2512 (val)
- 64.5 swiss-ai/Apertus-v1.1-1.5B (val)
- 64.4 NbAiLab/borealis2-26b-a4b-preview (zero-shot, val)
- 64.2 xai/grok-4.20-beta-0309-reasoning (zero-shot, val)
- 64.0 BSC-LT/ALIA-40b
- 64.0 deepseek/deepseek-flash#no-thinking (val)
- 63.7
Gemma 3 4B PT
- 63.6 google/gemma-4-12B-it (val)
- 63.5
Qwen3-4B-Thinking-2507
- 63.4 Qwen/Qwen3.5-35B-A3B-Base (val)
- 63.3 claude-sonnet-4-5-20250929#thinking (zero-shot, val)
- 62.9 NbAiLab/borealis2-26b-a4b-preview (val)
- 62.9 swiss-ai/Apertus-v1.1-0.5B (val)
- 62.7 ibm-granite/granite-4.0-h-tiny
- 62.4 mistralai/Mistral-Large-3-675B-Instruct-2512 (val)
- 62.3
Gemma 3 27B PT
- 62.3 gpt-4.1-mini-2025-04-14 (zero-shot, val)
- 62.1 allenai/Olmo-3.1-32B-Think (val)
- 62.0 Granite 4.0 H Small
- 61.9 LiquidAI/LFM2.5-8B-A1B-Base (val)
- 61.8 OpenLLM-Ro/RoMistral-7b-Instruct-DPO
- 61.7 mistralai/Mistral-Small-3.1-24B-Base-2503
- 61.3 OpenLLM-Ro/RoLlama3.1-8b-Instruct-DPO
- 61.2 typesafe/jev-1.13.0 (zero-shot)
- 61.1 amalia-llm/AMALIA-9B-0626-DPO (zero-shot, val)
- 61.1 gemini/gemini-2.5-flash-lite#no-thinking (zero-shot, val)
- 61.0 danish-foundation-models/DFM-Mimir-v1.8 (val)
- 60.9 swiss-ai/Apertus-v1.5-8B (zero-shot, val)
- 60.9 speakleash/Bielik-11B-v2.3-Instruct
- 60.8 allenai/Llama-3.1-Tulu-3-8B-SFT
- 60.5 utter-project/EuroLLM-1.7B-Instruct (val)
- 60.5 danish-foundation-models/DFM-Mimir-v1.8-Mini (val)
- 60.3 deepcogito/cogito-v1-preview-qwen-14B
- 60.0 mistralai/Mistral-Small-3.2-24B-Instruct-2506
- 59.9
Gemma 3 1B PT
- 59.7 gpt-oss-20b
- 59.6 amalia-llm/AMALIA-9B-0626-SFT (zero-shot, val)
- 59.6 BSC-LT/salamandra-7b-instruct
- 59.6 utter-project/EuroLLM-9B-Instruct-2512 (val)
- 59.5 syvai/claes-preview (zero-shot)
- 59.3 Qwen/Qwen3-1.7B#no-thinking
- 59.1 speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551 (zero-shot)
- 59.0 google/gemma-4-31B (val)
- 58.9 NbAiLab/borealis2-e4b-preview (zero-shot, val)
- 58.9 danish-foundation-models/DFM-Mimir (val)
- 58.9 deepcogito/cogito-v1-preview-qwen-32B
- 58.7 amalia-llm/AMALIA-9B-0626-DPO (val)
- 58.6 google/gemma-4-E4B-it (val)
- 58.6
Llama 3.3 70B Instruct
- 58.5 speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c (zero-shot)
- 58.4 meta-llama/Llama-3.1-8B-Instruct
- 58.4 Qwen/Qwen3-32B#no-thinking
- 58.4 Qwen/Qwen3-4B#no-thinking
- 58.4 HuggingFaceTB/SmolLM3-3B-Base
- 58.4 speakleash/Bielik-Minitron-7B-v3.0-Instruct (val)
- 58.0 mistralai/Mistral-Small-3.1-24B-Instruct-2503
- 57.6 allenai/Olmo-3.1-32B-Instruct-SFT
- 57.4 google/gemma-4-E2B-it (val)
- 57.3 gpt-4.1-nano-2025-04-14 (zero-shot, val)
- 57.3 LGAI-EXAONE/EXAONE-4.0-32B (val)
- 57.2 NbAiLab/borealis2-e2b-preview (zero-shot, val)
- 57.2 occiglot/occiglot-7b-eu5-instruct
- 56.9 NbAiLab/borealis2-e2b-preview (val)
- 56.5
Gemma 3 4B
- 56.4 danish-foundation-models/DFM-Mimir-v1.5 (val)
- 56.3 Qwen/Qwen3-14B#no-thinking
- 56.2 swiss-ai/Apertus-v1.5-8B (val)
- 56.1 Qwen/Qwen2.5-1.5B-Instruct
- 56.0 mistralai/Ministral-3-3B-Instruct-2512 (zero-shot, val)
- 55.9
Qwen2.5-72B
- 55.8 speakleash/Bielik-1.5B-v3.0-Instruct
- 55.8
Cogito v1 Preview 70B
- 55.8 speakleash/Bielik-11B-v3.0-Instruct (val)
- 55.8 Qwen/Qwen3-8B#no-thinking
- 55.7 OpenLLM-Ro/RoGemma2-9b-Instruct-DPO
- 55.6 kurakurai/Luth-LFM2-1.2B (val)
- 55.4 amalia-llm/AMALIA-9B-0626-SFT (val)
- 55.2 granite-4.0-micro
- 55.2
Voxtral Small
- 54.8 utter-project/EuroLLM-22B-Instruct-Preview
- 54.6 Unbabel/Tower-Plus-9B (val)
- 54.5 ibm-granite/granite-4.0-h-micro
- 54.5
Gemma 3 27B
- 54.5
Llama 3.2 1B
- 54.4 OpenLLM-France/Luciole-23B-Base (val)
- 54.4 deepcogito/cogito-v2-preview-llama-70B
- 53.8 NbAiLab/borealis2-e4b-preview (val)
- 53.6 speakleash/Bielik-11B-v3.1-Instruct (val)
- 53.5 google/rembert
- 53.3 baidu/ERNIE-4.5-21B-A3B-PT
- 53.2 deepcogito/cogito-v1-preview-llama-8B
- 53.2 yulan-team/YuLan-Mini-Instruct
- 53.2 danish-foundation-models/DFM-Mimir-v1.5 (zero-shot, val)
- 52.7 Qwen/Qwen3-0.6B#no-thinking
- 52.6 meta-llama/Llama-3.2-3B-Instruct
- 52.2 arcee-ai/AFM-4.5B-Base
- 52.2
Gemma 3 12B
- 51.9 mistralai/Ministral-3-3B-Instruct-2512 (val)
- 51.9 google/gemma-4-26B-A4B (val)
- 51.8 allenai/Olmo-3.1-32B-Instruct-DPO
- 51.2
Olmo 3.1 32B Instruct
- 51.1 LFM2-1.2B
- 51.1 setu4993/LaBSE
- 51.0 Aleph-Alpha/Kolibri-1 (zero-shot, val)
- 51.0 intfloat/multilingual-e5-base
- 50.6 Aleph-Alpha/Kolibri-1 (val)
- 50.3 jhu-clsp/mmBERT-base
- 50.3 mistralai/Ministral-3-8B-Reasoning-2512
- 50.2 speakleash/Bielik-4.5B-v3.0-Instruct
- 49.8 Unbabel/Tower-Plus-2B (val)
- 49.0 kurakurai/Luth-0.6B-Instruct
- 48.8 nvidia/NVIDIA-Nemotron-3-Nano-4B-FP8 (val)
- 48.8 LiquidAI/LFM2-2.6B
- 47.6
Gemma 3n E2B
- 46.7
OLMo 2 32B Instruct
- 46.6
Trinity Mini
- 46.3 swiss-ai/Apertus-v1.1-1.5B-Instruct (val)
- 46.2 sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2
- 46.1
Qwen3-0.6B
- 46.1 Qwen/Qwen3.5-0.8B (val)
- 46.0
Gemma 3n E4B
- 45.5 nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16
- 45.5 allenai/Olmo-3-1025-7B
- 44.9 deepcogito/cogito-v1-preview-llama-3B
- 44.9 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16#no-thinking
- 44.8 upstage/SOLAR-10.7B-v1.0
- 44.7 Qwen/Qwen3.5-9B-Base (val)
- 44.3
AFM-4.5B
- 44.3
Apertus 8B Instruct
- 44.1 swiss-ai/Apertus-v1.1-0.5B-Instruct (val)
- 44.1 utter-project/EuroLLM-9B
- 43.9
Ministral 3 3B Reasoning
- 43.8 Qwen/Qwen3.5-2B-Base (val)
- 43.6 LFM2.5-1.2B-Thinking
- 43.5 allenai/Olmo-3-7B-Instruct
- 43.4 utter-project/EuroLLM-9B-Instruct
- 42.9 PleIAs/Pleias-3b-Preview
- 42.6 allenai/Olmo-3-7B-Instruct-DPO
- 42.2 tencent/Hunyuan-7B-Instruct
- 42.1 allenai/Olmo-3-7B-Instruct-SFT
- 41.8 LiquidAI/LFM2.5-350M (val)
- 41.7 google/gemma-4-E4B (val)
- 41.7
Gemma 3 1B
- 41.5 LiquidAI/LFM2-700M
- 41.4 HuggingFaceTB/SmolLM3-3B#no-thinking
- 41.0 syvai/claes-v1 (zero-shot)
- 39.9 Qwen/Qwen3.5-4B-Base (val)
- 39.4 OpenLLM-France/Luciole-1B-Instruct-1.1 (val)
- 39.0 google/gemma-4-E2B (val)
- 38.7 fastino/GLiNER2.5-Decide (zero-shot)
- 36.3 google-bert/bert-base-multilingual-cased (val)
- 35.9 OmniJev/OneJev1.6-0.8B (zero-shot)
- 35.8 jhu-clsp/mmBERT-small
- 34.8 LFM2.5-1.2B-Instruct
- 33.8 kurakurai/Luth-LFM2-350M (val)
- 32.5 LiquidAI/LFM2.5-1.2B-Base
- 31.9 FacebookAI/xlm-roberta-large
- 29.8 microsoft/mdeberta-v3-base
- 29.4
SmolLM2-360M
- 29.0 PleIAs/Pleias-1.2b-Preview
- 28.4 AI-Sweden-Models/ModernBERT-large
- 27.4 utter-project/EuroMoE-2.6B-A0.6B-Preview
- 25.6 EuroBERT/EuroBERT-210m
- 25.4 arcee-ai/Trinity-Nano-Base
- 25.0 allenai/Olmo-3-1125-32B
- 24.1 convaiinnovations/laya (zero-shot)
- 23.7 AI-Sweden-Models/bert-large-nordic-pile-1M-steps
- 21.5 arcee-ai/Trinity-Nano-Preview
- 21.2 LiquidAI/LFM2-350M
- 21.0 EuroBERT/EuroBERT-610m
- 20.6 convaiinnovations/laya-multilingual (zero-shot)
- 19.7 jaredpalmer/kev-0.5b (zero-shot)
- 19.6 mii-llm/nesso-0.4B-agentic (val)
- 19.3 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview
- 18.9 answerdotai/ModernBERT-base
- 17.7 Qwen/Qwen3.5-0.8B-Base (val)
- 17.6
Gemma 3 270M
- 17.5 answerdotai/ModernBERT-large
- 17.4 EuropeanParliament/EUBERT
- 16.4 PleIAs/Pleias-350m-Preview
- 15.9 OpenLLM-France/Luciole-1B-Base (val)
- 15.8 speakleash/Bielik-11B-v3-Base-20250730
- 15.2 occiglot/occiglot-7b-eu5
- 15.2 BSC-LT/salamandra-7b
- 13.7 google-bert/bert-base-multilingual-uncased (val)
- 11.5 allenai/Olmo-3-7B-Think
- 10.6 fresh-xlm-roberta-base
- 10.6 baidu/ERNIE-4.5-0.3B-PT
- 9.84 HuggingFaceTB/SmolLM2-360M-Instruct
- 9.15 mii-llm/zagreus-0.4B-ita (val)
- 9.04 meta-llama/Llama-3.2-1B-Instruct
- 8.87
SmolLM2-135M
- 8.78 TildeAI/TildeOpen-30b (val)
- 8.64 distilbert/distilbert-base-multilingual-cased
- 7.08 LiquidAI/LFM2.5-8B-A1B (val)
- 6.13 openbmb/MiniCPM5-1B (val)
- 4.50 Geotrend/distilbert-base-25lang-cased
- 3.30 HuggingFaceTB/SmolLM2-135M-Instruct
- 3.04 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base
- 1.35 openGPT-X/Teuken-7B-instruct-v0.6 (val)
- 1.08 AIDC-AI/Marco-Mini-Global-Base (val)
- 0.99 inclusionAI/Ling-mini-base-2.0 (val)
- 0.57 inclusionAI/Ling-mini-2.0 (val)
- 0.51 PleIAs/Monad
- 0.25 Muse-research/Muse-3B (val)
- 0.14 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct
- 0.0 fresh-electra-small
- 0.0 openGPT-X/Teuken-7B-base-v0.6 (val)
- 0.0 PleIAs/Baguettotron@refs/pr/6#thinking
- 0.0 utter-project/EuroLLM-1.7B (val)
- 1 openeurollm/oellm-9b-128k-theta32m-v3 (val) val 78.3%
- 2 Qwen/Qwen3.5-9B (val) val 77.2%
- 3 Qwen/Qwen3.5-27B (val) val 76.9%
- 4 openai/gpt-5.4-mini-2026-03-17#high (zero-shot, val) zero-shot, val 76.1%
- 5 CohereLabs/tiny-aya-base (val) val 76%
- 6 zai-org/GLM-4.7-Flash (val) val 76%
- 7 gpt-5.6-sol (zero-shot, val) zero-shot, val 75.9%
- 8 CohereLabs/tiny-aya-global (val) val 75.2%
- 9 swiss-ai/Apertus-v1.1-4B (val) val 75.1%
- 10 CohereLabs/tiny-aya-water (val) val 74.4%
- 11 gemini/gemini-3.6-flash (zero-shot, val) zero-shot, val 74.1%
- 12 Qwen/Qwen3.6-27B-FP8 (val) val 74%
- 13 gemini/gemini-3-flash-preview#thinking (zero-shot, val) zero-shot, val 74%
- 14 gpt-5.6-terra (zero-shot, val) zero-shot, val 74%
- 15 gpt-5-mini-2025-08-07 (zero-shot, val) zero-shot, val 73.8%
- 16 openai/gpt-5.4-mini-2026-03-17#medium (zero-shot, val) zero-shot, val 73.8%
- 17 AI-Sweden-Models/Llama-3-8B-instruct (val) val 73.7%
- 18 gemini/gemini-3.7-flash (zero-shot, val) zero-shot, val 73.7%
- 19 speakleash/Bielik-11B-v2.3-Instruct (val) val 73.6%
- 20 gpt-6.2-sol (zero-shot, val) zero-shot, val 73.4%
- 21 Qwen/Qwen3.5-27B-FP8 (val) val 73.1%
- 22 Qwen/Qwen3.5-35B-A3B-FP8 (val) val 72.8%
- 23 Qwen/Qwen3.8-27B (val) val 72.7%
- 24 AI-Sweden-Models/Llama-3-8B-instruct (zero-shot, val) zero-shot, val 72.6%
- 25 Qwen/Qwen3.5-4B (val) val 72.5%
- 26 GadflyII/GLM-4.7-Flash-NVFP4 (val) val 72.5%
- 27 gpt-6-sol (zero-shot, val) zero-shot, val 72.4%
- 28 mistralai/Ministral-3-14B-Instruct-2512 (zero-shot, val) zero-shot, val 72.4%
- 29 openai/gpt-5.4-mini-2026-03-17#low (zero-shot, val) zero-shot, val 72.4%
- 30 google/gemma-4-31B-it (val) val 72.4%
- 31 gpt-4.1-2025-04-14 (zero-shot, val) zero-shot, val 72.4%
- 32 mistralai/Ministral-3-14B-Base-2512 (val) val 72.2%
- 33 claude-sonnet-4-6 (zero-shot, val) zero-shot, val 72.2%
- 34 gpt-6-astra (zero-shot, val) zero-shot, val 72%
- 35 OpenLLM-France/Lucie-7B-Instruct-v1.1 (val) val 72%
- 36 mistralai/Ministral-3-8B-Base-2512 (val) val 71.9%
- 37 mistralai/Ministral-3-14B-Reasoning-2512 (val) val 71.8%
- 38 swiss-ai/Apertus-8B-2509 71.8%
- 39 gpt-6.1-sol (zero-shot, val) zero-shot, val 71.7%
- 40 google/gemma-4-26B-A4B-it (val) val 71.5%
- 41 gpt-5-2025-08-07 (zero-shot, val) zero-shot, val 71.5%
- 42 gpt-6.2-astra (zero-shot, val) zero-shot, val 71.3%
- 43 swiss-ai/Apertus-v1.1-4B-Instruct (val) val 71.1%
- 44 mistralai/Ministral-3-8B-Instruct-2512 (val) val 71%
- 45 swiss-ai/Apertus-v1.5-70B (zero-shot, val) zero-shot, val 71%
- 46 latam-gpt/Llama-3.1-70B-LatamGPT-SFT-1.0 (val) val 70.7%
- 47 openai/gpt-5.4-nano-2026-03-17#low (zero-shot, val) zero-shot, val 70.7%
- 48 openeurollm/prelude-base (val) val 70.7%
- 49 mistralai/Mistral-Small-4-119B-2603 (zero-shot, val) zero-shot, val 70.5%
- 50 gpt-6-luna (zero-shot, val) zero-shot, val 70.3%
- 51 mistralai/Mistral-Medium-3.5-128B (zero-shot, val) zero-shot, val 70.3%
- 52 CohereLabs/aya-23-8B (val) val 70.2%
- 53 deepseek/deepseek-flash#low (zero-shot, val) zero-shot, val 69.7%
- 54 zai-org/GLM-5.3-Flash (val) val 69.5%
- 55
Apertus 70B 69.5%
- 56
Llama 3.1-8B 69.5%
- 57 Qwen/Qwen3.5-2B (val) val 69.4%
- 58 domyn/Domyn-Small-v1.0 (val) val 69.4%
- 59 openai/gpt-5.4-nano-2026-03-17#medium (zero-shot, val) zero-shot, val 69.3%
- 60 gemini/gemini-3.5-flash-lite (zero-shot, val) zero-shot, val 69%
- 61 mistralai/Mistral-Small-4-119B-2603 (val) val 68.9%
- 62 gemini/gemini-2.5-pro (zero-shot, val) zero-shot, val 68.7%
- 63 gemini/gemini-2.5-flash-lite#thinking (zero-shot, val) zero-shot, val 68.6%
- 64 openai/gpt-5.4-mini-2026-03-17#none (zero-shot, val) zero-shot, val 68.6%
- 65 gemini/gemini-2.5-flash#thinking (zero-shot, val) zero-shot, val 68.5%
- 66 OpenLLM-France/Luciole-8B-Base (val) val 68.5%
- 67 NbAiLab/borealis2-31b-preview (zero-shot, val) zero-shot, val 68.5%
- 68 claude-haiku-4-5-20251001 (zero-shot, val) zero-shot, val 68.5%
- 69 openai/gpt-5.4-2026-03-05#none (zero-shot, val) zero-shot, val 68.4%
- 70 gpt-5-nano-2025-08-07 (zero-shot, val) zero-shot, val 68.3%
- 71 utter-project/EuroLLM-22B-2512 68.3%
- 72 Unbabel/Tower-Plus-72B (val) val 68.2%
- 73 claude-sonnet-4-5-20250929#no-thinking (zero-shot, val) zero-shot, val 68.1%
- 74 OpenLLM-France/Luciole-8B-Instruct-1.1 (val) val 68.1%
- 75 openai/gpt-5.4-nano-2026-03-17#none (zero-shot, val) zero-shot, val 68.1%
- 76 openai/gpt-5.6-luna (zero-shot, val) zero-shot, val 68%
- 77 xai/grok-4-1-fast-reasoning (zero-shot, val) zero-shot, val 67.8%
- 78 gemini/gemini-3.1-flash-lite-preview (zero-shot, val) zero-shot, val 67.7%
- 79
Gemma 3 12B PT 67.5%
- 80 openeurollm/prelude-long (val) val 67.4%
- 81 NbAiLab/borealis2-31b-preview (val) val 67.3%
- 82 nvidia/Nemotron-Cascade-2-30B-A3B (val) val 67.3%
- 83 gemini/gemini-3-pro-preview (zero-shot, val) zero-shot, val 67.3%
- 84 OpenLLM-France/Luciole-23B-Instruct-1.1 (val) val 67.2%
- 85
Llama 3.1-70B 67.2%
- 86 gpt-5.2-2025-12-11 (zero-shot, val) zero-shot, val 67.2%
- 87 swiss-ai/Apertus-v1.5-70B (val) val 67.1%
- 88 mistralai/Mistral-Medium-3.5-128B (val) val 67.1%
- 89 utter-project/EuroLLM-22B-Instruct-2512 67.1%
- 90 openai/gpt-5.4-nano-2026-03-17#high (zero-shot, val) zero-shot, val 67%
- 91 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 (val) val 66.9%
- 92 google/gemma-4-12B (val) val 66.7%
- 93 mii-llm/nesso-4B 66.7%
- 94 zai-org/GLM-5.3-Flash (zero-shot, val) zero-shot, val 66.5%
- 95 TrustLLMeu/baseline-7-8b_2-3t-tokens_llama 66.5%
- 96 CohereLabs/aya-expanse-8b (val) val 66.5%
- 97 mistralai/Mistral-Large-3-675B-Instruct-2512 (zero-shot, val) zero-shot, val 66.2%
- 98 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8 (val) val 66.2%
- 99 gemini/gemini-2.5-flash#no-thinking (zero-shot, val) zero-shot, val 66%
- 100 gemini/gemini-3-flash-preview#no-thinking (zero-shot, val) zero-shot, val 66%
- 101 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-Base-BF16 (val) val 66%
- 102 danish-foundation-models/DFM-Mimir-v1.8-Mini (zero-shot, val) zero-shot, val 66%
- 103 OpenLLM-France/Lucie-7B-Instruct-human-data (val) val 65.9%
- 104 intfloat/multilingual-e5-large 65.8%
- 105 openai/gpt-oss-120b (val) val 65.5%
- 106 mistralai/Ministral-3-3B-Base-2512 65.5%
- 107 danish-foundation-models/DFM-Mimir-v1.8 (zero-shot, val) zero-shot, val 65.5%
- 108 utter-project/EuroLLM-22B-Preview 65.5%
- 109 danish-foundation-models/DFM-Mimir (zero-shot, val) zero-shot, val 65.3%
- 110 speakleash/Bielik-11B-v2.6-Instruct 65.3%
- 111 mistralai/Ministral-3-14B-Instruct-2512 (val) val 65.1%
- 112 swiss-ai/Apertus-v1.1-1.5B (val) val 64.5%
- 113 NbAiLab/borealis2-26b-a4b-preview (zero-shot, val) zero-shot, val 64.4%
- 114 xai/grok-4.20-beta-0309-reasoning (zero-shot, val) zero-shot, val 64.2%
- 115 BSC-LT/ALIA-40b 64%
- 116 deepseek/deepseek-flash#no-thinking (val) val 64%
- 117
Gemma 3 4B PT 63.7%
- 118 google/gemma-4-12B-it (val) val 63.6%
- 119
Qwen3-4B-Thinking-2507 63.5%
- 120 Qwen/Qwen3.5-35B-A3B-Base (val) val 63.4%
- 121 claude-sonnet-4-5-20250929#thinking (zero-shot, val) zero-shot, val 63.3%
- 122 NbAiLab/borealis2-26b-a4b-preview (val) val 62.9%
- 123 swiss-ai/Apertus-v1.1-0.5B (val) val 62.9%
- 124 ibm-granite/granite-4.0-h-tiny 62.7%
- 125 mistralai/Mistral-Large-3-675B-Instruct-2512 (val) val 62.4%
- 126
Gemma 3 27B PT 62.3%
- 127 gpt-4.1-mini-2025-04-14 (zero-shot, val) zero-shot, val 62.3%
- 128 allenai/Olmo-3.1-32B-Think (val) val 62.1%
- 129 Granite 4.0 H Small 62%
- 130 LiquidAI/LFM2.5-8B-A1B-Base (val) val 61.9%
- 131 OpenLLM-Ro/RoMistral-7b-Instruct-DPO 61.8%
- 132 mistralai/Mistral-Small-3.1-24B-Base-2503 61.7%
- 133 OpenLLM-Ro/RoLlama3.1-8b-Instruct-DPO 61.3%
- 134 typesafe/jev-1.13.0 (zero-shot) zero-shot 61.2%
- 135 amalia-llm/AMALIA-9B-0626-DPO (zero-shot, val) zero-shot, val 61.1%
- 136 gemini/gemini-2.5-flash-lite#no-thinking (zero-shot, val) zero-shot, val 61.1%
- 137 danish-foundation-models/DFM-Mimir-v1.8 (val) val 61%
- 138 swiss-ai/Apertus-v1.5-8B (zero-shot, val) zero-shot, val 60.9%
- 139 speakleash/Bielik-11B-v2.3-Instruct 60.9%
- 140 allenai/Llama-3.1-Tulu-3-8B-SFT 60.8%
- 141 utter-project/EuroLLM-1.7B-Instruct (val) val 60.5%
- 142 danish-foundation-models/DFM-Mimir-v1.8-Mini (val) val 60.5%
- 143 deepcogito/cogito-v1-preview-qwen-14B 60.3%
- 144 mistralai/Mistral-Small-3.2-24B-Instruct-2506 60%
- 145
Gemma 3 1B PT 59.9%
- 146 gpt-oss-20b medium 59.7%
- 147 amalia-llm/AMALIA-9B-0626-SFT (zero-shot, val) zero-shot, val 59.6%
- 148 BSC-LT/salamandra-7b-instruct 59.6%
- 149 utter-project/EuroLLM-9B-Instruct-2512 (val) val 59.6%
- 150 syvai/claes-preview (zero-shot) zero-shot 59.5%
- 151 Qwen/Qwen3-1.7B#no-thinking 59.3%
- 152 speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551 (zero-shot) zero-shot 59.1%
- 153 google/gemma-4-31B (val) val 59%
- 154 NbAiLab/borealis2-e4b-preview (zero-shot, val) zero-shot, val 58.9%
- 155 danish-foundation-models/DFM-Mimir (val) val 58.9%
- 156 deepcogito/cogito-v1-preview-qwen-32B 58.9%
- 157 amalia-llm/AMALIA-9B-0626-DPO (val) val 58.7%
- 158 google/gemma-4-E4B-it (val) val 58.6%
- 159
Llama 3.3 70B Instruct 58.6%
- 160 speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c (zero-shot) zero-shot 58.5%
- 161 meta-llama/Llama-3.1-8B-Instruct 58.4%
- 162 Qwen/Qwen3-32B#no-thinking 58.4%
- 163 Qwen/Qwen3-4B#no-thinking 58.4%
- 164 HuggingFaceTB/SmolLM3-3B-Base 58.4%
- 165 speakleash/Bielik-Minitron-7B-v3.0-Instruct (val) val 58.4%
- 166 mistralai/Mistral-Small-3.1-24B-Instruct-2503 58%
- 167 allenai/Olmo-3.1-32B-Instruct-SFT 57.6%
- 168 google/gemma-4-E2B-it (val) val 57.4%
- 169 gpt-4.1-nano-2025-04-14 (zero-shot, val) zero-shot, val 57.3%
- 170 LGAI-EXAONE/EXAONE-4.0-32B (val) val 57.3%
- 171 NbAiLab/borealis2-e2b-preview (zero-shot, val) zero-shot, val 57.2%
- 172 occiglot/occiglot-7b-eu5-instruct 57.2%
- 173 NbAiLab/borealis2-e2b-preview (val) val 56.9%
- 174
Gemma 3 4B 56.5%
- 175 danish-foundation-models/DFM-Mimir-v1.5 (val) val 56.4%
- 176 Qwen/Qwen3-14B#no-thinking 56.3%
- 177 swiss-ai/Apertus-v1.5-8B (val) val 56.2%
- 178 Qwen/Qwen2.5-1.5B-Instruct 56.1%
- 179 mistralai/Ministral-3-3B-Instruct-2512 (zero-shot, val) zero-shot, val 56%
- 180
Qwen2.5-72B 55.9%
- 181 speakleash/Bielik-1.5B-v3.0-Instruct 55.8%
- 182
Cogito v1 Preview 70B 55.8%
- 183 speakleash/Bielik-11B-v3.0-Instruct (val) val 55.8%
- 184 Qwen/Qwen3-8B#no-thinking 55.8%
- 185 OpenLLM-Ro/RoGemma2-9b-Instruct-DPO 55.7%
- 186 kurakurai/Luth-LFM2-1.2B (val) val 55.6%
- 187 amalia-llm/AMALIA-9B-0626-SFT (val) val 55.4%
- 188 granite-4.0-micro 55.2%
- 189
Voxtral Small 55.2%
- 190 utter-project/EuroLLM-22B-Instruct-Preview 54.8%
- 191 Unbabel/Tower-Plus-9B (val) val 54.6%
- 192 ibm-granite/granite-4.0-h-micro 54.5%
- 193
Gemma 3 27B 54.5%
- 194
Llama 3.2 1B 54.5%
- 195 OpenLLM-France/Luciole-23B-Base (val) val 54.4%
- 196 deepcogito/cogito-v2-preview-llama-70B 54.4%
- 197 NbAiLab/borealis2-e4b-preview (val) val 53.8%
- 198 speakleash/Bielik-11B-v3.1-Instruct (val) val 53.6%
- 199 google/rembert 53.5%
- 200 baidu/ERNIE-4.5-21B-A3B-PT 53.3%
- 201 deepcogito/cogito-v1-preview-llama-8B 53.2%
- 202 yulan-team/YuLan-Mini-Instruct 53.2%
- 203 danish-foundation-models/DFM-Mimir-v1.5 (zero-shot, val) zero-shot, val 53.2%
- 204 Qwen/Qwen3-0.6B#no-thinking 52.7%
- 205 meta-llama/Llama-3.2-3B-Instruct 52.6%
- 206 arcee-ai/AFM-4.5B-Base 52.2%
- 207
Gemma 3 12B 52.2%
- 208 mistralai/Ministral-3-3B-Instruct-2512 (val) val 51.9%
- 209 google/gemma-4-26B-A4B (val) val 51.9%
- 210 allenai/Olmo-3.1-32B-Instruct-DPO 51.8%
- 211
Olmo 3.1 32B Instruct 51.2%
- 212 LFM2-1.2B 51.1%
- 213 setu4993/LaBSE 51.1%
- 214 Aleph-Alpha/Kolibri-1 (zero-shot, val) zero-shot, val 51%
- 215 intfloat/multilingual-e5-base 51%
- 216 Aleph-Alpha/Kolibri-1 (val) val 50.6%
- 217 jhu-clsp/mmBERT-base 50.3%
- 218 mistralai/Ministral-3-8B-Reasoning-2512 50.3%
- 219 speakleash/Bielik-4.5B-v3.0-Instruct 50.2%
- 220 Unbabel/Tower-Plus-2B (val) val 49.8%
- 221 kurakurai/Luth-0.6B-Instruct 49%
- 222 nvidia/NVIDIA-Nemotron-3-Nano-4B-FP8 (val) val 48.8%
- 223 LiquidAI/LFM2-2.6B 48.8%
- 224
Gemma 3n E2B 47.6%
- 225
OLMo 2 32B Instruct 46.7%
- 226
Trinity Mini 46.6%
- 227 swiss-ai/Apertus-v1.1-1.5B-Instruct (val) val 46.3%
- 228 sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 46.2%
- 229
Qwen3-0.6B 46.1%
- 230 Qwen/Qwen3.5-0.8B (val) val 46.1%
- 231
Gemma 3n E4B 46%
- 232 nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16 45.5%
- 233 allenai/Olmo-3-1025-7B 45.5%
- 234 deepcogito/cogito-v1-preview-llama-3B 44.9%
- 235 nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16#no-thinking 44.9%
- 236 upstage/SOLAR-10.7B-v1.0 44.8%
- 237 Qwen/Qwen3.5-9B-Base (val) val 44.7%
- 238
AFM-4.5B 44.3%
- 239
Apertus 8B Instruct 44.3%
- 240 swiss-ai/Apertus-v1.1-0.5B-Instruct (val) val 44.1%
- 241 utter-project/EuroLLM-9B 44.1%
- 242
Ministral 3 3B Reasoning 43.9%
- 243 Qwen/Qwen3.5-2B-Base (val) val 43.8%
- 244 LFM2.5-1.2B-Thinking 43.6%
- 245 allenai/Olmo-3-7B-Instruct 43.5%
- 246 utter-project/EuroLLM-9B-Instruct 43.4%
- 247 PleIAs/Pleias-3b-Preview 42.9%
- 248 allenai/Olmo-3-7B-Instruct-DPO 42.6%
- 249 tencent/Hunyuan-7B-Instruct 42.2%
- 250 allenai/Olmo-3-7B-Instruct-SFT 42.1%
- 251 LiquidAI/LFM2.5-350M (val) val 41.8%
- 252 google/gemma-4-E4B (val) val 41.7%
- 253
Gemma 3 1B 41.7%
- 254 LiquidAI/LFM2-700M 41.5%
- 255 HuggingFaceTB/SmolLM3-3B#no-thinking 41.4%
- 256 syvai/claes-v1 (zero-shot) zero-shot 41%
- 257 Qwen/Qwen3.5-4B-Base (val) val 39.9%
- 258 OpenLLM-France/Luciole-1B-Instruct-1.1 (val) val 39.4%
- 259 google/gemma-4-E2B (val) val 39%
- 260 fastino/GLiNER2.5-Decide (zero-shot) zero-shot 38.7%
- 261 google-bert/bert-base-multilingual-cased (val) val 36.3%
- 262 OmniJev/OneJev1.6-0.8B (zero-shot) zero-shot 35.9%
- 263 jhu-clsp/mmBERT-small 35.8%
- 264 LFM2.5-1.2B-Instruct 34.8%
- 265 kurakurai/Luth-LFM2-350M (val) val 33.8%
- 266 LiquidAI/LFM2.5-1.2B-Base 32.5%
- 267 FacebookAI/xlm-roberta-large 31.9%
- 268 microsoft/mdeberta-v3-base 29.8%
- 269
SmolLM2-360M 29.4%
- 270 PleIAs/Pleias-1.2b-Preview 29%
- 271 AI-Sweden-Models/ModernBERT-large 28.4%
- 272 utter-project/EuroMoE-2.6B-A0.6B-Preview 27.4%
- 273 EuroBERT/EuroBERT-210m 25.6%
- 274 arcee-ai/Trinity-Nano-Base 25.4%
- 275 allenai/Olmo-3-1125-32B 25%
- 276 convaiinnovations/laya (zero-shot) zero-shot 24.1%
- 277 AI-Sweden-Models/bert-large-nordic-pile-1M-steps 23.7%
- 278 arcee-ai/Trinity-Nano-Preview 21.5%
- 279 LiquidAI/LFM2-350M 21.2%
- 280 EuroBERT/EuroBERT-610m 21%
- 281 convaiinnovations/laya-multilingual (zero-shot) zero-shot 20.6%
- 282 jaredpalmer/kev-0.5b (zero-shot) zero-shot 19.7%
- 283 mii-llm/nesso-0.4B-agentic (val) val 19.6%
- 284 utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview 19.3%
- 285 answerdotai/ModernBERT-base 18.9%
- 286 Qwen/Qwen3.5-0.8B-Base (val) val 17.7%
- 287
Gemma 3 270M 17.6%
- 288 answerdotai/ModernBERT-large 17.5%
- 289 EuropeanParliament/EUBERT 17.4%
- 290 PleIAs/Pleias-350m-Preview 16.4%
- 291 OpenLLM-France/Luciole-1B-Base (val) val 15.9%
- 292 speakleash/Bielik-11B-v3-Base-20250730 15.8%
- 293 occiglot/occiglot-7b-eu5 15.2%
- 294 BSC-LT/salamandra-7b 15.2%
- 295 google-bert/bert-base-multilingual-uncased (val) val 13.7%
- 296 allenai/Olmo-3-7B-Think 11.5%
- 297 fresh-xlm-roberta-base 10.6%
- 298 baidu/ERNIE-4.5-0.3B-PT 10.6%
- 299 HuggingFaceTB/SmolLM2-360M-Instruct 9.84%
- 300 mii-llm/zagreus-0.4B-ita (val) val 9.15%
- 301 meta-llama/Llama-3.2-1B-Instruct 9.04%
- 302
SmolLM2-135M 8.87%
- 303 TildeAI/TildeOpen-30b (val) val 8.78%
- 304 distilbert/distilbert-base-multilingual-cased 8.64%
- 305 LiquidAI/LFM2.5-8B-A1B (val) val 7.08%
- 306 openbmb/MiniCPM5-1B (val) val 6.13%
- 307 Geotrend/distilbert-base-25lang-cased 4.50%
- 308 HuggingFaceTB/SmolLM2-135M-Instruct 3.30%
- 309 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base 3.04%
- 310 openGPT-X/Teuken-7B-instruct-v0.6 (val) val 1.35%
- 311 AIDC-AI/Marco-Mini-Global-Base (val) val 1.08%
- 312 inclusionAI/Ling-mini-base-2.0 (val) val 0.99%
- 313 inclusionAI/Ling-mini-2.0 (val) val 0.57%
- 314 PleIAs/Monad 0.51%
- 315 Muse-research/Muse-3B (val) val 0.25%
- 316 tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct 0.14%
- 317 fresh-electra-small 0%
- 318 openGPT-X/Teuken-7B-base-v0.6 (val) val 0%
- 319 PleIAs/Baguettotron@refs/pr/6#thinking 0%
- 320 utter-project/EuroLLM-1.7B (val) val 0%
Source: Official GuiaCat leaderboard
- 69.5
Apertus 70B
- 69.5
Llama 3.1-8B
- 67.5
Gemma 3 12B PT
- 67.2
Llama 3.1-70B
- 63.7
Gemma 3 4B PT
- 63.5
Qwen3-4B-Thinking-2507
- 62.3
Gemma 3 27B PT
- 62.0 Granite 4.0 H Small
- 59.9
Gemma 3 1B PT
- 59.7 gpt-oss-20b
- 58.6
Llama 3.3 70B Instruct
- 56.5
Gemma 3 4B
- 55.9
Qwen2.5-72B
- 55.8
Cogito v1 Preview 70B
- 55.2 granite-4.0-micro
- 55.2
Voxtral Small
- 54.5
Gemma 3 27B
- 54.5
Llama 3.2 1B
- 52.2
Gemma 3 12B
- 51.2
Olmo 3.1 32B Instruct
- 51.1 LFM2-1.2B
- 47.6
Gemma 3n E2B
- 46.7
OLMo 2 32B Instruct
- 46.6
Trinity Mini
- 46.1
Qwen3-0.6B
- 46.0
Gemma 3n E4B
- 44.3
AFM-4.5B
- 44.3
Apertus 8B Instruct
- 43.9
Ministral 3 3B Reasoning
- 43.6 LFM2.5-1.2B-Thinking
- 41.7
Gemma 3 1B
- 34.8 LFM2.5-1.2B-Instruct
- 29.4
SmolLM2-360M
- 17.6
Gemma 3 270M
- 8.87
SmolLM2-135M
- 1
Apertus 70B 69.5%
- 2
Llama 3.1-8B 69.5%
- 3
Gemma 3 12B PT 67.5%
- 4
Llama 3.1-70B 67.2%
- 5
Gemma 3 4B PT 63.7%
- 6
Qwen3-4B-Thinking-2507 63.5%
- 7
Gemma 3 27B PT 62.3%
- 8 Granite 4.0 H Small 62%
- 9
Gemma 3 1B PT 59.9%
- 10 gpt-oss-20b medium 59.7%
- 11
Llama 3.3 70B Instruct 58.6%
- 12
Gemma 3 4B 56.5%
- 13
Qwen2.5-72B 55.9%
- 14
Cogito v1 Preview 70B 55.8%
- 15 granite-4.0-micro 55.2%
- 16
Voxtral Small 55.2%
- 17
Gemma 3 27B 54.5%
- 18
Llama 3.2 1B 54.5%
- 19
Gemma 3 12B 52.2%
- 20
Olmo 3.1 32B Instruct 51.2%
- 21 LFM2-1.2B 51.1%
- 22
Gemma 3n E2B 47.6%
- 23
OLMo 2 32B Instruct 46.7%
- 24
Trinity Mini 46.6%
- 25
Qwen3-0.6B 46.1%
- 26
Gemma 3n E4B 46%
- 27
AFM-4.5B 44.3%
- 28
Apertus 8B Instruct 44.3%
- 29
Ministral 3 3B Reasoning 43.9%
- 30 LFM2.5-1.2B-Thinking 43.6%
- 31
Gemma 3 1B 41.7%
- 32 LFM2.5-1.2B-Instruct 34.8%
- 33
SmolLM2-360M 29.4%
- 34
Gemma 3 270M 17.6%
- 35
SmolLM2-135M 8.87%
Source: Official GuiaCat leaderboard
| # | Entry | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | openeurollm/oellm-9b-128k-theta32m-v3 (val) | 78.3% | val | Official leaderboard | |
| 2 | Qwen/Qwen3.5-9B (val) | 77.2% | val | Official leaderboard | |
| 3 | Qwen/Qwen3.5-27B (val) | 76.9% | val | Official leaderboard | |
| 4 | openai/gpt-5.4-mini-2026-03-17#high (zero-shot, val) | 76.1% | zero-shot, val | Official leaderboard | |
| 5 | CohereLabs/tiny-aya-base (val) | 76% | val | Official leaderboard | |
| 6 | zai-org/GLM-4.7-Flash (val) | 76% | val | Official leaderboard | |
| 7 | gpt-5.6-sol (zero-shot, val) | 75.9% | zero-shot, val | Official leaderboard | |
| 8 | CohereLabs/tiny-aya-global (val) | 75.2% | val | Official leaderboard | |
| 9 | swiss-ai/Apertus-v1.1-4B (val) | 75.1% | val | Official leaderboard | |
| 10 | CohereLabs/tiny-aya-water (val) | 74.4% | val | Official leaderboard | |
| 11 | gemini/gemini-3.6-flash (zero-shot, val) | 74.1% | zero-shot, val | Official leaderboard | |
| 12 | Qwen/Qwen3.6-27B-FP8 (val) | 74% | val | Official leaderboard | |
| 13 | gemini/gemini-3-flash-preview#thinking (zero-shot, val) | 74% | zero-shot, val | Official leaderboard | |
| 14 | gpt-5.6-terra (zero-shot, val) | 74% | zero-shot, val | Official leaderboard | |
| 15 | gpt-5-mini-2025-08-07 (zero-shot, val) | 73.8% | zero-shot, val | Official leaderboard | |
| 16 | openai/gpt-5.4-mini-2026-03-17#medium (zero-shot, val) | 73.8% | zero-shot, val | Official leaderboard | |
| 17 | AI-Sweden-Models/Llama-3-8B-instruct (val) | 73.7% | val | Official leaderboard | |
| 18 | gemini/gemini-3.7-flash (zero-shot, val) | 73.7% | zero-shot, val | Official leaderboard | |
| 19 | speakleash/Bielik-11B-v2.3-Instruct (val) | 73.6% | val | Official leaderboard | |
| 20 | gpt-6.2-sol (zero-shot, val) | 73.4% | zero-shot, val | Official leaderboard | |
| 21 | Qwen/Qwen3.5-27B-FP8 (val) | 73.1% | val | Official leaderboard | |
| 22 | Qwen/Qwen3.5-35B-A3B-FP8 (val) | 72.8% | val | Official leaderboard | |
| 23 | Qwen/Qwen3.8-27B (val) | 72.7% | val | Official leaderboard | |
| 24 | AI-Sweden-Models/Llama-3-8B-instruct (zero-shot, val) | 72.6% | zero-shot, val | Official leaderboard | |
| 25 | Qwen/Qwen3.5-4B (val) | 72.5% | val | Official leaderboard | |
| 26 | GadflyII/GLM-4.7-Flash-NVFP4 (val) | 72.5% | val | Official leaderboard | |
| 27 | gpt-6-sol (zero-shot, val) | 72.4% | zero-shot, val | Official leaderboard | |
| 28 | mistralai/Ministral-3-14B-Instruct-2512 (zero-shot, val) | 72.4% | zero-shot, val | Official leaderboard | |
| 29 | openai/gpt-5.4-mini-2026-03-17#low (zero-shot, val) | 72.4% | zero-shot, val | Official leaderboard | |
| 30 | google/gemma-4-31B-it (val) | 72.4% | val | Official leaderboard | |
| 31 | gpt-4.1-2025-04-14 (zero-shot, val) | 72.4% | zero-shot, val | Official leaderboard | |
| 32 | mistralai/Ministral-3-14B-Base-2512 (val) | 72.2% | val | Official leaderboard | |
| 33 | claude-sonnet-4-6 (zero-shot, val) | 72.2% | zero-shot, val | Official leaderboard | |
| 34 | gpt-6-astra (zero-shot, val) | 72% | zero-shot, val | Official leaderboard | |
| 35 | OpenLLM-France/Lucie-7B-Instruct-v1.1 (val) | 72% | val | Official leaderboard | |
| 36 | mistralai/Ministral-3-8B-Base-2512 (val) | 71.9% | val | Official leaderboard | |
| 37 | mistralai/Ministral-3-14B-Reasoning-2512 (val) | 71.8% | val | Official leaderboard | |
| 38 | swiss-ai/Apertus-8B-2509 | 71.8% | Official leaderboard | ||
| 39 | gpt-6.1-sol (zero-shot, val) | 71.7% | zero-shot, val | Official leaderboard | |
| 40 | google/gemma-4-26B-A4B-it (val) | 71.5% | val | Official leaderboard | |
| 41 | gpt-5-2025-08-07 (zero-shot, val) | 71.5% | zero-shot, val | Official leaderboard | |
| 42 | gpt-6.2-astra (zero-shot, val) | 71.3% | zero-shot, val | Official leaderboard | |
| 43 | swiss-ai/Apertus-v1.1-4B-Instruct (val) | 71.1% | val | Official leaderboard | |
| 44 | mistralai/Ministral-3-8B-Instruct-2512 (val) | 71% | val | Official leaderboard | |
| 45 | swiss-ai/Apertus-v1.5-70B (zero-shot, val) | 71% | zero-shot, val | Official leaderboard | |
| 46 | latam-gpt/Llama-3.1-70B-LatamGPT-SFT-1.0 (val) | 70.7% | val | Official leaderboard | |
| 47 | openai/gpt-5.4-nano-2026-03-17#low (zero-shot, val) | 70.7% | zero-shot, val | Official leaderboard | |
| 48 | openeurollm/prelude-base (val) | 70.7% | val | Official leaderboard | |
| 49 | mistralai/Mistral-Small-4-119B-2603 (zero-shot, val) | 70.5% | zero-shot, val | Official leaderboard | |
| 50 | gpt-6-luna (zero-shot, val) | 70.3% | zero-shot, val | Official leaderboard | |
| 51 | mistralai/Mistral-Medium-3.5-128B (zero-shot, val) | 70.3% | zero-shot, val | Official leaderboard | |
| 52 | CohereLabs/aya-23-8B (val) | 70.2% | val | Official leaderboard | |
| 53 | deepseek/deepseek-flash#low (zero-shot, val) | 69.7% | zero-shot, val | Official leaderboard | |
| 54 | zai-org/GLM-5.3-Flash (val) | 69.5% | val | Official leaderboard | |
| 55 | | 69.5% | Sep 2025 | Official leaderboard | |
| 56 | | 69.5% | Jul 2024 | Official leaderboard | |
| 57 | Qwen/Qwen3.5-2B (val) | 69.4% | val | Official leaderboard | |
| 58 | domyn/Domyn-Small-v1.0 (val) | 69.4% | val | Official leaderboard | |
| 59 | openai/gpt-5.4-nano-2026-03-17#medium (zero-shot, val) | 69.3% | zero-shot, val | Official leaderboard | |
| 60 | gemini/gemini-3.5-flash-lite (zero-shot, val) | 69% | zero-shot, val | Official leaderboard | |
| 61 | mistralai/Mistral-Small-4-119B-2603 (val) | 68.9% | val | Official leaderboard | |
| 62 | gemini/gemini-2.5-pro (zero-shot, val) | 68.7% | zero-shot, val | Official leaderboard | |
| 63 | gemini/gemini-2.5-flash-lite#thinking (zero-shot, val) | 68.6% | zero-shot, val | Official leaderboard | |
| 64 | openai/gpt-5.4-mini-2026-03-17#none (zero-shot, val) | 68.6% | zero-shot, val | Official leaderboard | |
| 65 | gemini/gemini-2.5-flash#thinking (zero-shot, val) | 68.5% | zero-shot, val | Official leaderboard | |
| 66 | OpenLLM-France/Luciole-8B-Base (val) | 68.5% | val | Official leaderboard | |
| 67 | NbAiLab/borealis2-31b-preview (zero-shot, val) | 68.5% | zero-shot, val | Official leaderboard | |
| 68 | claude-haiku-4-5-20251001 (zero-shot, val) | 68.5% | zero-shot, val | Official leaderboard | |
| 69 | openai/gpt-5.4-2026-03-05#none (zero-shot, val) | 68.4% | zero-shot, val | Official leaderboard | |
| 70 | gpt-5-nano-2025-08-07 (zero-shot, val) | 68.3% | zero-shot, val | Official leaderboard | |
| 71 | utter-project/EuroLLM-22B-2512 | 68.3% | Official leaderboard | ||
| 72 | Unbabel/Tower-Plus-72B (val) | 68.2% | val | Official leaderboard | |
| 73 | claude-sonnet-4-5-20250929#no-thinking (zero-shot, val) | 68.1% | zero-shot, val | Official leaderboard | |
| 74 | OpenLLM-France/Luciole-8B-Instruct-1.1 (val) | 68.1% | val | Official leaderboard | |
| 75 | openai/gpt-5.4-nano-2026-03-17#none (zero-shot, val) | 68.1% | zero-shot, val | Official leaderboard | |
| 76 | openai/gpt-5.6-luna (zero-shot, val) | 68% | zero-shot, val | Official leaderboard | |
| 77 | xai/grok-4-1-fast-reasoning (zero-shot, val) | 67.8% | zero-shot, val | Official leaderboard | |
| 78 | gemini/gemini-3.1-flash-lite-preview (zero-shot, val) | 67.7% | zero-shot, val | Official leaderboard | |
| 79 | | 67.5% | Mar 2025 | Official leaderboard | |
| 80 | openeurollm/prelude-long (val) | 67.4% | val | Official leaderboard | |
| 81 | NbAiLab/borealis2-31b-preview (val) | 67.3% | val | Official leaderboard | |
| 82 | nvidia/Nemotron-Cascade-2-30B-A3B (val) | 67.3% | val | Official leaderboard | |
| 83 | gemini/gemini-3-pro-preview (zero-shot, val) | 67.3% | zero-shot, val | Official leaderboard | |
| 84 | OpenLLM-France/Luciole-23B-Instruct-1.1 (val) | 67.2% | val | Official leaderboard | |
| 85 | | 67.2% | Jul 2024 | Official leaderboard | |
| 86 | gpt-5.2-2025-12-11 (zero-shot, val) | 67.2% | zero-shot, val | Official leaderboard | |
| 87 | swiss-ai/Apertus-v1.5-70B (val) | 67.1% | val | Official leaderboard | |
| 88 | mistralai/Mistral-Medium-3.5-128B (val) | 67.1% | val | Official leaderboard | |
| 89 | utter-project/EuroLLM-22B-Instruct-2512 | 67.1% | Official leaderboard | ||
| 90 | openai/gpt-5.4-nano-2026-03-17#high (zero-shot, val) | 67% | zero-shot, val | Official leaderboard | |
| 91 | nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 (val) | 66.9% | val | Official leaderboard | |
| 92 | google/gemma-4-12B (val) | 66.7% | val | Official leaderboard | |
| 93 | mii-llm/nesso-4B | 66.7% | Official leaderboard | ||
| 94 | zai-org/GLM-5.3-Flash (zero-shot, val) | 66.5% | zero-shot, val | Official leaderboard | |
| 95 | TrustLLMeu/baseline-7-8b_2-3t-tokens_llama | 66.5% | Official leaderboard | ||
| 96 | CohereLabs/aya-expanse-8b (val) | 66.5% | val | Official leaderboard | |
| 97 | mistralai/Mistral-Large-3-675B-Instruct-2512 (zero-shot, val) | 66.2% | zero-shot, val | Official leaderboard | |
| 98 | nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-FP8 (val) | 66.2% | val | Official leaderboard | |
| 99 | gemini/gemini-2.5-flash#no-thinking (zero-shot, val) | 66% | zero-shot, val | Official leaderboard | |
| 100 | gemini/gemini-3-flash-preview#no-thinking (zero-shot, val) | 66% | zero-shot, val | Official leaderboard | |
| 101 | nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-Base-BF16 (val) | 66% | val | Official leaderboard | |
| 102 | danish-foundation-models/DFM-Mimir-v1.8-Mini (zero-shot, val) | 66% | zero-shot, val | Official leaderboard | |
| 103 | OpenLLM-France/Lucie-7B-Instruct-human-data (val) | 65.9% | val | Official leaderboard | |
| 104 | intfloat/multilingual-e5-large | 65.8% | Official leaderboard | ||
| 105 | openai/gpt-oss-120b (val) | 65.5% | val | Official leaderboard | |
| 106 | mistralai/Ministral-3-3B-Base-2512 | 65.5% | Official leaderboard | ||
| 107 | danish-foundation-models/DFM-Mimir-v1.8 (zero-shot, val) | 65.5% | zero-shot, val | Official leaderboard | |
| 108 | utter-project/EuroLLM-22B-Preview | 65.5% | Official leaderboard | ||
| 109 | danish-foundation-models/DFM-Mimir (zero-shot, val) | 65.3% | zero-shot, val | Official leaderboard | |
| 110 | speakleash/Bielik-11B-v2.6-Instruct | 65.3% | Official leaderboard | ||
| 111 | mistralai/Ministral-3-14B-Instruct-2512 (val) | 65.1% | val | Official leaderboard | |
| 112 | swiss-ai/Apertus-v1.1-1.5B (val) | 64.5% | val | Official leaderboard | |
| 113 | NbAiLab/borealis2-26b-a4b-preview (zero-shot, val) | 64.4% | zero-shot, val | Official leaderboard | |
| 114 | xai/grok-4.20-beta-0309-reasoning (zero-shot, val) | 64.2% | zero-shot, val | Official leaderboard | |
| 115 | BSC-LT/ALIA-40b | 64% | Official leaderboard | ||
| 116 | deepseek/deepseek-flash#no-thinking (val) | 64% | val | Official leaderboard | |
| 117 | | 63.7% | Mar 2025 | Official leaderboard | |
| 118 | google/gemma-4-12B-it (val) | 63.6% | val | Official leaderboard | |
| 119 | | 63.5% | Aug 2025 | Official leaderboard | |
| 120 | Qwen/Qwen3.5-35B-A3B-Base (val) | 63.4% | val | Official leaderboard | |
| 121 | claude-sonnet-4-5-20250929#thinking (zero-shot, val) | 63.3% | zero-shot, val | Official leaderboard | |
| 122 | NbAiLab/borealis2-26b-a4b-preview (val) | 62.9% | val | Official leaderboard | |
| 123 | swiss-ai/Apertus-v1.1-0.5B (val) | 62.9% | val | Official leaderboard | |
| 124 | ibm-granite/granite-4.0-h-tiny | 62.7% | Official leaderboard | ||
| 125 | mistralai/Mistral-Large-3-675B-Instruct-2512 (val) | 62.4% | val | Official leaderboard | |
| 126 | | 62.3% | Mar 2025 | Official leaderboard | |
| 127 | gpt-4.1-mini-2025-04-14 (zero-shot, val) | 62.3% | zero-shot, val | Official leaderboard | |
| 128 | allenai/Olmo-3.1-32B-Think (val) | 62.1% | val | Official leaderboard | |
| 129 | Granite 4.0 H Small | 62% | Oct 2025 | Official leaderboard | |
| 130 | LiquidAI/LFM2.5-8B-A1B-Base (val) | 61.9% | val | Official leaderboard | |
| 131 | OpenLLM-Ro/RoMistral-7b-Instruct-DPO | 61.8% | Official leaderboard | ||
| 132 | mistralai/Mistral-Small-3.1-24B-Base-2503 | 61.7% | Official leaderboard | ||
| 133 | OpenLLM-Ro/RoLlama3.1-8b-Instruct-DPO | 61.3% | Official leaderboard | ||
| 134 | typesafe/jev-1.13.0 (zero-shot) | 61.2% | zero-shot | Official leaderboard | |
| 135 | amalia-llm/AMALIA-9B-0626-DPO (zero-shot, val) | 61.1% | zero-shot, val | Official leaderboard | |
| 136 | gemini/gemini-2.5-flash-lite#no-thinking (zero-shot, val) | 61.1% | zero-shot, val | Official leaderboard | |
| 137 | danish-foundation-models/DFM-Mimir-v1.8 (val) | 61% | val | Official leaderboard | |
| 138 | swiss-ai/Apertus-v1.5-8B (zero-shot, val) | 60.9% | zero-shot, val | Official leaderboard | |
| 139 | speakleash/Bielik-11B-v2.3-Instruct | 60.9% | Official leaderboard | ||
| 140 | allenai/Llama-3.1-Tulu-3-8B-SFT | 60.8% | Official leaderboard | ||
| 141 | utter-project/EuroLLM-1.7B-Instruct (val) | 60.5% | val | Official leaderboard | |
| 142 | danish-foundation-models/DFM-Mimir-v1.8-Mini (val) | 60.5% | val | Official leaderboard | |
| 143 | deepcogito/cogito-v1-preview-qwen-14B | 60.3% | Official leaderboard | ||
| 144 | mistralai/Mistral-Small-3.2-24B-Instruct-2506 | 60% | Official leaderboard | ||
| 145 | | 59.9% | Mar 2025 | Official leaderboard | |
| 146 | gpt-oss-20b | 59.7% | medium | Aug 2025 | Official leaderboard |
| 147 | amalia-llm/AMALIA-9B-0626-SFT (zero-shot, val) | 59.6% | zero-shot, val | Official leaderboard | |
| 148 | BSC-LT/salamandra-7b-instruct | 59.6% | Official leaderboard | ||
| 149 | utter-project/EuroLLM-9B-Instruct-2512 (val) | 59.6% | val | Official leaderboard | |
| 150 | syvai/claes-preview (zero-shot) | 59.5% | zero-shot | Official leaderboard | |
| 151 | Qwen/Qwen3-1.7B#no-thinking | 59.3% | Official leaderboard | ||
| 152 | speakleash/Bielik-One-4B-r1@567d29b0f15df28794acf0f6187f9216860d4551 (zero-shot) | 59.1% | zero-shot | Official leaderboard | |
| 153 | google/gemma-4-31B (val) | 59% | val | Official leaderboard | |
| 154 | NbAiLab/borealis2-e4b-preview (zero-shot, val) | 58.9% | zero-shot, val | Official leaderboard | |
| 155 | danish-foundation-models/DFM-Mimir (val) | 58.9% | val | Official leaderboard | |
| 156 | deepcogito/cogito-v1-preview-qwen-32B | 58.9% | Official leaderboard | ||
| 157 | amalia-llm/AMALIA-9B-0626-DPO (val) | 58.7% | val | Official leaderboard | |
| 158 | google/gemma-4-E4B-it (val) | 58.6% | val | Official leaderboard | |
| 159 | | 58.6% | Dec 2024 | Official leaderboard | |
| 160 | speakleash/Bielik-One-11B-v3.1-r1@75a2d83890e3c68bf7fac2fe5bbe2bff6372036c (zero-shot) | 58.5% | zero-shot | Official leaderboard | |
| 161 | meta-llama/Llama-3.1-8B-Instruct | 58.4% | Official leaderboard | ||
| 162 | Qwen/Qwen3-32B#no-thinking | 58.4% | Official leaderboard | ||
| 163 | Qwen/Qwen3-4B#no-thinking | 58.4% | Official leaderboard | ||
| 164 | HuggingFaceTB/SmolLM3-3B-Base | 58.4% | Official leaderboard | ||
| 165 | speakleash/Bielik-Minitron-7B-v3.0-Instruct (val) | 58.4% | val | Official leaderboard | |
| 166 | mistralai/Mistral-Small-3.1-24B-Instruct-2503 | 58% | Official leaderboard | ||
| 167 | allenai/Olmo-3.1-32B-Instruct-SFT | 57.6% | Official leaderboard | ||
| 168 | google/gemma-4-E2B-it (val) | 57.4% | val | Official leaderboard | |
| 169 | gpt-4.1-nano-2025-04-14 (zero-shot, val) | 57.3% | zero-shot, val | Official leaderboard | |
| 170 | LGAI-EXAONE/EXAONE-4.0-32B (val) | 57.3% | val | Official leaderboard | |
| 171 | NbAiLab/borealis2-e2b-preview (zero-shot, val) | 57.2% | zero-shot, val | Official leaderboard | |
| 172 | occiglot/occiglot-7b-eu5-instruct | 57.2% | Official leaderboard | ||
| 173 | NbAiLab/borealis2-e2b-preview (val) | 56.9% | val | Official leaderboard | |
| 174 | | 56.5% | Mar 2025 | Official leaderboard | |
| 175 | danish-foundation-models/DFM-Mimir-v1.5 (val) | 56.4% | val | Official leaderboard | |
| 176 | Qwen/Qwen3-14B#no-thinking | 56.3% | Official leaderboard | ||
| 177 | swiss-ai/Apertus-v1.5-8B (val) | 56.2% | val | Official leaderboard | |
| 178 | Qwen/Qwen2.5-1.5B-Instruct | 56.1% | Official leaderboard | ||
| 179 | mistralai/Ministral-3-3B-Instruct-2512 (zero-shot, val) | 56% | zero-shot, val | Official leaderboard | |
| 180 | | 55.9% | Sep 2024 | Official leaderboard | |
| 181 | speakleash/Bielik-1.5B-v3.0-Instruct | 55.8% | Official leaderboard | ||
| 182 | | 55.8% | Apr 2025 | Official leaderboard | |
| 183 | speakleash/Bielik-11B-v3.0-Instruct (val) | 55.8% | val | Official leaderboard | |
| 184 | Qwen/Qwen3-8B#no-thinking | 55.8% | Official leaderboard | ||
| 185 | OpenLLM-Ro/RoGemma2-9b-Instruct-DPO | 55.7% | Official leaderboard | ||
| 186 | kurakurai/Luth-LFM2-1.2B (val) | 55.6% | val | Official leaderboard | |
| 187 | amalia-llm/AMALIA-9B-0626-SFT (val) | 55.4% | val | Official leaderboard | |
| 188 | granite-4.0-micro | 55.2% | Oct 2025 | Official leaderboard | |
| 189 | | 55.2% | Jul 2025 | Official leaderboard | |
| 190 | utter-project/EuroLLM-22B-Instruct-Preview | 54.8% | Official leaderboard | ||
| 191 | Unbabel/Tower-Plus-9B (val) | 54.6% | val | Official leaderboard | |
| 192 | ibm-granite/granite-4.0-h-micro | 54.5% | Official leaderboard | ||
| 193 | | 54.5% | Mar 2025 | Official leaderboard | |
| 194 | | 54.5% | Sep 2024 | Official leaderboard | |
| 195 | OpenLLM-France/Luciole-23B-Base (val) | 54.4% | val | Official leaderboard | |
| 196 | deepcogito/cogito-v2-preview-llama-70B | 54.4% | Official leaderboard | ||
| 197 | NbAiLab/borealis2-e4b-preview (val) | 53.8% | val | Official leaderboard | |
| 198 | speakleash/Bielik-11B-v3.1-Instruct (val) | 53.6% | val | Official leaderboard | |
| 199 | google/rembert | 53.5% | Official leaderboard | ||
| 200 | baidu/ERNIE-4.5-21B-A3B-PT | 53.3% | Official leaderboard | ||
| 201 | deepcogito/cogito-v1-preview-llama-8B | 53.2% | Official leaderboard | ||
| 202 | yulan-team/YuLan-Mini-Instruct | 53.2% | Official leaderboard | ||
| 203 | danish-foundation-models/DFM-Mimir-v1.5 (zero-shot, val) | 53.2% | zero-shot, val | Official leaderboard | |
| 204 | Qwen/Qwen3-0.6B#no-thinking | 52.7% | Official leaderboard | ||
| 205 | meta-llama/Llama-3.2-3B-Instruct | 52.6% | Official leaderboard | ||
| 206 | arcee-ai/AFM-4.5B-Base | 52.2% | Official leaderboard | ||
| 207 | | 52.2% | Mar 2025 | Official leaderboard | |
| 208 | mistralai/Ministral-3-3B-Instruct-2512 (val) | 51.9% | val | Official leaderboard | |
| 209 | google/gemma-4-26B-A4B (val) | 51.9% | val | Official leaderboard | |
| 210 | allenai/Olmo-3.1-32B-Instruct-DPO | 51.8% | Official leaderboard | ||
| 211 | | 51.2% | Dec 2025 | Official leaderboard | |
| 212 | LFM2-1.2B | 51.1% | Jul 2025 | Official leaderboard | |
| 213 | setu4993/LaBSE | 51.1% | Official leaderboard | ||
| 214 | Aleph-Alpha/Kolibri-1 (zero-shot, val) | 51% | zero-shot, val | Official leaderboard | |
| 215 | intfloat/multilingual-e5-base | 51% | Official leaderboard | ||
| 216 | Aleph-Alpha/Kolibri-1 (val) | 50.6% | val | Official leaderboard | |
| 217 | jhu-clsp/mmBERT-base | 50.3% | Official leaderboard | ||
| 218 | mistralai/Ministral-3-8B-Reasoning-2512 | 50.3% | Official leaderboard | ||
| 219 | speakleash/Bielik-4.5B-v3.0-Instruct | 50.2% | Official leaderboard | ||
| 220 | Unbabel/Tower-Plus-2B (val) | 49.8% | val | Official leaderboard | |
| 221 | kurakurai/Luth-0.6B-Instruct | 49% | Official leaderboard | ||
| 222 | nvidia/NVIDIA-Nemotron-3-Nano-4B-FP8 (val) | 48.8% | val | Official leaderboard | |
| 223 | LiquidAI/LFM2-2.6B | 48.8% | Official leaderboard | ||
| 224 | | 47.6% | Jun 2025 | Official leaderboard | |
| 225 | | 46.7% | Mar 2025 | Official leaderboard | |
| 226 | | 46.6% | Dec 2025 | Official leaderboard | |
| 227 | swiss-ai/Apertus-v1.1-1.5B-Instruct (val) | 46.3% | val | Official leaderboard | |
| 228 | sentence-transformers/paraphrase-multilingual-MiniLM-L12-v2 | 46.2% | Official leaderboard | ||
| 229 | | 46.1% | Apr 2025 | Official leaderboard | |
| 230 | Qwen/Qwen3.5-0.8B (val) | 46.1% | val | Official leaderboard | |
| 231 | | 46% | Jun 2025 | Official leaderboard | |
| 232 | nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16 | 45.5% | Official leaderboard | ||
| 233 | allenai/Olmo-3-1025-7B | 45.5% | Official leaderboard | ||
| 234 | deepcogito/cogito-v1-preview-llama-3B | 44.9% | Official leaderboard | ||
| 235 | nvidia/NVIDIA-Nemotron-3-Nano-30B-A3B-BF16#no-thinking | 44.9% | Official leaderboard | ||
| 236 | upstage/SOLAR-10.7B-v1.0 | 44.8% | Official leaderboard | ||
| 237 | Qwen/Qwen3.5-9B-Base (val) | 44.7% | val | Official leaderboard | |
| 238 | | 44.3% | Jul 2025 | Official leaderboard | |
| 239 | | 44.3% | Sep 2025 | Official leaderboard | |
| 240 | swiss-ai/Apertus-v1.1-0.5B-Instruct (val) | 44.1% | val | Official leaderboard | |
| 241 | utter-project/EuroLLM-9B | 44.1% | Official leaderboard | ||
| 242 | | 43.9% | Dec 2025 | Official leaderboard | |
| 243 | Qwen/Qwen3.5-2B-Base (val) | 43.8% | val | Official leaderboard | |
| 244 | LFM2.5-1.2B-Thinking | 43.6% | Jan 2026 | Official leaderboard | |
| 245 | allenai/Olmo-3-7B-Instruct | 43.5% | Official leaderboard | ||
| 246 | utter-project/EuroLLM-9B-Instruct | 43.4% | Official leaderboard | ||
| 247 | PleIAs/Pleias-3b-Preview | 42.9% | Official leaderboard | ||
| 248 | allenai/Olmo-3-7B-Instruct-DPO | 42.6% | Official leaderboard | ||
| 249 | tencent/Hunyuan-7B-Instruct | 42.2% | Official leaderboard | ||
| 250 | allenai/Olmo-3-7B-Instruct-SFT | 42.1% | Official leaderboard | ||
| 251 | LiquidAI/LFM2.5-350M (val) | 41.8% | val | Official leaderboard | |
| 252 | google/gemma-4-E4B (val) | 41.7% | val | Official leaderboard | |
| 253 | | 41.7% | Mar 2025 | Official leaderboard | |
| 254 | LiquidAI/LFM2-700M | 41.5% | Official leaderboard | ||
| 255 | HuggingFaceTB/SmolLM3-3B#no-thinking | 41.4% | Official leaderboard | ||
| 256 | syvai/claes-v1 (zero-shot) | 41% | zero-shot | Official leaderboard | |
| 257 | Qwen/Qwen3.5-4B-Base (val) | 39.9% | val | Official leaderboard | |
| 258 | OpenLLM-France/Luciole-1B-Instruct-1.1 (val) | 39.4% | val | Official leaderboard | |
| 259 | google/gemma-4-E2B (val) | 39% | val | Official leaderboard | |
| 260 | fastino/GLiNER2.5-Decide (zero-shot) | 38.7% | zero-shot | Official leaderboard | |
| 261 | google-bert/bert-base-multilingual-cased (val) | 36.3% | val | Official leaderboard | |
| 262 | OmniJev/OneJev1.6-0.8B (zero-shot) | 35.9% | zero-shot | Official leaderboard | |
| 263 | jhu-clsp/mmBERT-small | 35.8% | Official leaderboard | ||
| 264 | LFM2.5-1.2B-Instruct | 34.8% | Jan 2026 | Official leaderboard | |
| 265 | kurakurai/Luth-LFM2-350M (val) | 33.8% | val | Official leaderboard | |
| 266 | LiquidAI/LFM2.5-1.2B-Base | 32.5% | Official leaderboard | ||
| 267 | FacebookAI/xlm-roberta-large | 31.9% | Official leaderboard | ||
| 268 | microsoft/mdeberta-v3-base | 29.8% | Official leaderboard | ||
| 269 | | 29.4% | Oct 2024 | Official leaderboard | |
| 270 | PleIAs/Pleias-1.2b-Preview | 29% | Official leaderboard | ||
| 271 | AI-Sweden-Models/ModernBERT-large | 28.4% | Official leaderboard | ||
| 272 | utter-project/EuroMoE-2.6B-A0.6B-Preview | 27.4% | Official leaderboard | ||
| 273 | EuroBERT/EuroBERT-210m | 25.6% | Official leaderboard | ||
| 274 | arcee-ai/Trinity-Nano-Base | 25.4% | Official leaderboard | ||
| 275 | allenai/Olmo-3-1125-32B | 25% | Official leaderboard | ||
| 276 | convaiinnovations/laya (zero-shot) | 24.1% | zero-shot | Official leaderboard | |
| 277 | AI-Sweden-Models/bert-large-nordic-pile-1M-steps | 23.7% | Official leaderboard | ||
| 278 | arcee-ai/Trinity-Nano-Preview | 21.5% | Official leaderboard | ||
| 279 | LiquidAI/LFM2-350M | 21.2% | Official leaderboard | ||
| 280 | EuroBERT/EuroBERT-610m | 21% | Official leaderboard | ||
| 281 | convaiinnovations/laya-multilingual (zero-shot) | 20.6% | zero-shot | Official leaderboard | |
| 282 | jaredpalmer/kev-0.5b (zero-shot) | 19.7% | zero-shot | Official leaderboard | |
| 283 | mii-llm/nesso-0.4B-agentic (val) | 19.6% | val | Official leaderboard | |
| 284 | utter-project/EuroMoE-2.6B-A0.6B-Instruct-Preview | 19.3% | Official leaderboard | ||
| 285 | answerdotai/ModernBERT-base | 18.9% | Official leaderboard | ||
| 286 | Qwen/Qwen3.5-0.8B-Base (val) | 17.7% | val | Official leaderboard | |
| 287 | | 17.6% | Aug 2025 | Official leaderboard | |
| 288 | answerdotai/ModernBERT-large | 17.5% | Official leaderboard | ||
| 289 | EuropeanParliament/EUBERT | 17.4% | Official leaderboard | ||
| 290 | PleIAs/Pleias-350m-Preview | 16.4% | Official leaderboard | ||
| 291 | OpenLLM-France/Luciole-1B-Base (val) | 15.9% | val | Official leaderboard | |
| 292 | speakleash/Bielik-11B-v3-Base-20250730 | 15.8% | Official leaderboard | ||
| 293 | occiglot/occiglot-7b-eu5 | 15.2% | Official leaderboard | ||
| 294 | BSC-LT/salamandra-7b | 15.2% | Official leaderboard | ||
| 295 | google-bert/bert-base-multilingual-uncased (val) | 13.7% | val | Official leaderboard | |
| 296 | allenai/Olmo-3-7B-Think | 11.5% | Official leaderboard | ||
| 297 | fresh-xlm-roberta-base | 10.6% | Official leaderboard | ||
| 298 | baidu/ERNIE-4.5-0.3B-PT | 10.6% | Official leaderboard | ||
| 299 | HuggingFaceTB/SmolLM2-360M-Instruct | 9.84% | Official leaderboard | ||
| 300 | mii-llm/zagreus-0.4B-ita (val) | 9.15% | val | Official leaderboard | |
| 301 | meta-llama/Llama-3.2-1B-Instruct | 9.04% | Official leaderboard | ||
| 302 | | 8.87% | Oct 2024 | Official leaderboard | |
| 303 | TildeAI/TildeOpen-30b (val) | 8.78% | val | Official leaderboard | |
| 304 | distilbert/distilbert-base-multilingual-cased | 8.64% | Official leaderboard | ||
| 305 | LiquidAI/LFM2.5-8B-A1B (val) | 7.08% | val | Official leaderboard | |
| 306 | openbmb/MiniCPM5-1B (val) | 6.13% | val | Official leaderboard | |
| 307 | Geotrend/distilbert-base-25lang-cased | 4.50% | Official leaderboard | ||
| 308 | HuggingFaceTB/SmolLM2-135M-Instruct | 3.30% | Official leaderboard | ||
| 309 | tiiuae/Falcon-H1-Tiny-Multilingual-100M-Base | 3.04% | Official leaderboard | ||
| 310 | openGPT-X/Teuken-7B-instruct-v0.6 (val) | 1.35% | val | Official leaderboard | |
| 311 | AIDC-AI/Marco-Mini-Global-Base (val) | 1.08% | val | Official leaderboard | |
| 312 | inclusionAI/Ling-mini-base-2.0 (val) | 0.99% | val | Official leaderboard | |
| 313 | inclusionAI/Ling-mini-2.0 (val) | 0.57% | val | Official leaderboard | |
| 314 | PleIAs/Monad | 0.51% | Official leaderboard | ||
| 315 | Muse-research/Muse-3B (val) | 0.25% | val | Official leaderboard | |
| 316 | tiiuae/Falcon-H1-Tiny-Multilingual-100M-Instruct | 0.14% | Official leaderboard | ||
| 317 | fresh-electra-small | 0% | Official leaderboard | ||
| 318 | openGPT-X/Teuken-7B-base-v0.6 (val) | 0% | val | Official leaderboard | |
| 319 | PleIAs/Baguettotron@refs/pr/6#thinking | 0% | Official leaderboard | ||
| 320 | utter-project/EuroLLM-1.7B (val) | 0% | val | Official leaderboard |
| # | Model | Score | Setting | Date | Source |
|---|---|---|---|---|---|
| 1 | | 69.5% | Sep 2025 | Official leaderboard | |
| 2 | | 69.5% | Jul 2024 | Official leaderboard | |
| 3 | | 67.5% | Mar 2025 | Official leaderboard | |
| 4 | | 67.2% | Jul 2024 | Official leaderboard | |
| 5 | | 63.7% | Mar 2025 | Official leaderboard | |
| 6 | | 63.5% | Aug 2025 | Official leaderboard | |
| 7 | | 62.3% | Mar 2025 | Official leaderboard | |
| 8 | Granite 4.0 H Small | 62% | Oct 2025 | Official leaderboard | |
| 9 | | 59.9% | Mar 2025 | Official leaderboard | |
| 10 | gpt-oss-20b | 59.7% | medium | Aug 2025 | Official leaderboard |
| 11 | | 58.6% | Dec 2024 | Official leaderboard | |
| 12 | | 56.5% | Mar 2025 | Official leaderboard | |
| 13 | | 55.9% | Sep 2024 | Official leaderboard | |
| 14 | | 55.8% | Apr 2025 | Official leaderboard | |
| 15 | granite-4.0-micro | 55.2% | Oct 2025 | Official leaderboard | |
| 16 | | 55.2% | Jul 2025 | Official leaderboard | |
| 17 | | 54.5% | Mar 2025 | Official leaderboard | |
| 18 | | 54.5% | Sep 2024 | Official leaderboard | |
| 19 | | 52.2% | Mar 2025 | Official leaderboard | |
| 20 | | 51.2% | Dec 2025 | Official leaderboard | |
| 21 | LFM2-1.2B | 51.1% | Jul 2025 | Official leaderboard | |
| 22 | | 47.6% | Jun 2025 | Official leaderboard | |
| 23 | | 46.7% | Mar 2025 | Official leaderboard | |
| 24 | | 46.6% | Dec 2025 | Official leaderboard | |
| 25 | | 46.1% | Apr 2025 | Official leaderboard | |
| 26 | | 46% | Jun 2025 | Official leaderboard | |
| 27 | | 44.3% | Jul 2025 | Official leaderboard | |
| 28 | | 44.3% | Sep 2025 | Official leaderboard | |
| 29 | | 43.9% | Dec 2025 | Official leaderboard | |
| 30 | LFM2.5-1.2B-Thinking | 43.6% | Jan 2026 | Official leaderboard | |
| 31 | | 41.7% | Mar 2025 | Official leaderboard | |
| 32 | LFM2.5-1.2B-Instruct | 34.8% | Jan 2026 | Official leaderboard | |
| 33 | | 29.4% | Oct 2024 | Official leaderboard | |
| 34 | | 17.6% | Aug 2025 | Official leaderboard | |
| 35 | | 8.87% | Oct 2024 | Official leaderboard |
2 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.
Source: Official GuiaCat 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 GuiaCat 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 10 months ago, in January 2026, for LFM2.5-1.2B-Thinking. 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 (320 entries).
- Pass Models scored 35 models we track, plus 285 other entries on its leaderboard.
- Warning Newest score The newest came 10 months ago, in January 2026, for LFM2.5-1.2B-Thinking.
- Note Frontier models None of the top 10 has a score. The strongest tested is gpt-oss-20b (#143).
- Pass Headroom The best is 69.5%, under 85%, so it still separates models.
Leaderboard import 3 pass
- Pass Importer We read the board on euroeval.com: 39 rows for models we track and 285 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.