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

GPQA

46 models

GPQA (Graduate-Level Google-Proof Q&A Benchmark) measures an AI's ability to answer graduate-level, highly challenging questions in biology, physics, and chemistry that are designed to be "Google-proof" — requiring deep domain expertise and reasoning rather than simple lookup.

Top 5 Models Performance

Google: Gemini 3.1 Pro Preview ######################################## 94.3
Anthropic: Claude Opus 4.7 ######################################## 94.2
OpenAI: GPT-5.5 ######################################## 93.6
Z.ai: GLM 5.2 ####################################### 91.2
Moonshot AI: Kimi K2.6 ###################################### 90.5
– 970 – 1.1T
– 2019 – 2026
Rank Model Score
🇺🇸 1 Google: Gemini 3.1 Pro Preview 94.3
🇺🇸 2 Anthropic: Claude Opus 4.7 94.2
🇺🇸 3 OpenAI: GPT-5.5 93.6
🇨🇳 4 Z.ai: GLM 5.2北京智谱华章科技有限公司 91.2
🇨🇳 5 Moonshot AI: Kimi K2.6月之暗面 90.5
🇨🇳 6 DeepSeek: deepseek-v4-pro-0813深度求索 90.1
🇨🇳 7 DeepSeek: DeepSeek V4 Pro深度求索 90.1
🇨🇳 8 DeepSeek: DeepSeek V4 Flash深度求索 88.1
🇺🇸 9 OpenAI: gpt-5.5-instant 85.6
🇨🇳 10 Moonshot AI: Kimi K2 Thinking月之暗面 84.5
🇨🇳 11 DeepSeek: deepseek-v3.2深度求索 79.9
🇨🇳 12 Intern Language Models: intern-s1 77.3
🇺🇸 13 Inception: Mercury 2 74
🇺🇸 14 NVIDIA Corporation: nvidia-nemotron-3-nano-30b-a3b 73
❓ 15 xhtoken: spark-x2.6-4b 69.9
🇺🇸 16 AMD: sand-math-qwen2.5-32b 68.72
🇺🇸 17 AMD: sand-mathscience-deepseek-qwen32b 68.72
❓ 18 xhtoken: spark-x2.5-4b 67.4
🇨🇳 19 Xiaomi: MiMo-V2.5 Pro小米集团 66.7
🇺🇸 20 NVIDIA Corporation: nvidia-nemotron-nano-9b-v2 64
🇨🇳 21 Qwen: Qwen3 8B通义千问 59.6
🇺🇸 22 OpenAI: gpt-4o 53.6
🇺🇸 23 NVIDIA Corporation: NVIDIA Nemotron 3 Nano 4B 53.2
🇷🇺 24 Yandex: gpt-5.1-proЯндекс 46
🇫🇷 25 Mistral AI: mistral-small-3.2-24b-instruct-2506 44.22
❓ 26 xhtoken: spark-x2.5-1.7b 43.8
🇷🇺 27 Yandex: gpt-5-proЯндекс 42
🇦🇪 28 Technology Innovation Institute: falcon-h1-1.5b-deep-instruct 33.22
🇦🇪 29 Technology Innovation Institute: falcon-h1-1.5b-deep-base 32.8
🇺🇸 30 OpenAI: GPT-3.5 Turbo 30.8
❓ 31 zyphra: zaya1-base 30.7
🇺🇸 32 Meta Llama: Llama 3.2 3B Instruct 30.6
🇺🇸 33 Google: Gemma 3 4B Instruct 29.51
🇺🇸 34 LiquidAI: lfm-2-8b-a1b 29.29
🇦🇪 35 Technology Innovation Institute: falcon-h1-1.5b-base 29.11
🇨🇳 36 Qwen: qwen3-1.7b-base通义千问 28.28
🇨🇳 37 Qwen: qwen3-0.6b-base通义千问 26.77
🇦🇪 38 Technology Innovation Institute: falcon-h1-1.5b-instruct 26.34
🇺🇸 39 Google: gemma-3-12b-pt 25.4
🇨🇳 40 Qwen: qwen2.5-0.5b通义千问 24.75
🇺🇸 41 Google: gemma-3-1b-pt 24.75
🇺🇸 42 Google: Gemma 3 27B Pretrained 24.3
🇨🇳 43 Qwen: qwen2.5-1.5b通义千问 24.24
🇺🇸 44 Google: gemma-3-4b-pt 15
🇨🇳 45 Qwen: Qwen3.5 0.8B通义千问 11.9
🇦🇪 46 Technology Innovation Institute: falcon3-mamba-7b-base 10.6