Note: Overall leaderboard rankings may not reflect true model quality — individual benchmarks give a clearer picture. ARC-Challenge MMLU GPQA GSM8K Artificial Analysis Intelligence Index v4.1
← Back to leaderboard

BBH

33 models

Top 5 Models Performance

Xiaomi: MiMo-V2.5 Pro ######################################## 88.4
Sber: gigachat3.5-432b-a28b-base ######################################## 87.5
DeepSeek: deepseek-v2.5 ###################################### 84.3
Qwen: Qwen3 4B ################################### 77.8
Tencent: youtu-llm-2b ################################### 77.5
– 970 – 1.1T
– 2019 – 2026
Rank Model Score
🇨🇳 1 Xiaomi: MiMo-V2.5 Pro小米集团 88.4
🇷🇺 2 Sber: gigachat3.5-432b-a28b-baseСбер 87.5
🇨🇳 3 DeepSeek: deepseek-v2.5深度求索 84.3
🇨🇳 4 Qwen: Qwen3 4B通义千问 77.8
🇨🇳 5 Tencent: youtu-llm-2b腾讯控股有限公司 77.5
🇺🇸 6 Hugging Face Smol Models Research: smollm3-3b 76.3
🇷🇺 7 Yandex: gpt-5-lite-pretrainЯндекс 73.91
❓ 8 cloudflare: clef 73.7
🇨🇳 9 OpenBMB: minicpm5-1b 71.89
🇨🇳 10 Qwen: qwen3-1.7b通义千问 69.1
❓ 11 cloudflare: clef-flash 68.9
🇨🇳 12 Moonshot AI: Moonlight Instruct月之暗面 65.2
🇨🇳 13 OpenBMB: bitcpm4-1b 57.64
🇨🇳 14 Qwen: qwen3-1.7b-base通义千问 54.47
🇦🇪 15 Technology Innovation Institute: falcon-h1-1.5b-deep-instruct 54.43
🇨🇳 16 Qwen: qwen-14b通义千问 53.4
🇦🇪 17 Technology Innovation Institute: falcon-h1-1.5b-deep-base 52.37
🇺🇸 18 Meta Llama: llama-2-70b-chat 51.2
🇺🇸 19 Arcee AI: trinity-nano-base 50.4
🇦🇪 20 Technology Innovation Institute: falcon-h1-1.5b-base 46.57
🇦🇪 21 Technology Innovation Institute: falcon-h1-1.5b-instruct 46.47
🇨🇳 22 Qwen: qwen2.5-1.5b通义千问 45.1
🇨🇳 23 Qwen: qwen-7b通义千问 45
🇨🇳 24 Qwen: qwen3-0.6b-base通义千问 41.47
🇺🇸 25 Meta Llama: llama-2-13b-chat 39.4
🇺🇸 26 Meta Llama: llama-2-7b-chat 32.6
🇨🇳 27 DeepSeek: deepseek-r1-distill-qwen-1.5b深度求索 31
🇨🇳 28 Qwen: qwen-1.8b通义千问 28.2
🇺🇸 29 Hugging Face Smol Models Research: smollm2-135m-instruct 28.2
🇺🇸 30 Google: gemma-3-1b-pt 28.13
🇦🇪 31 Technology Innovation Institute: falcon3-mamba-7b-base 25.6
🇺🇸 32 AMD: amd-olmo-1b 20.95
🇨🇳 33 Qwen: qwen2.5-0.5b通义千问 20.3