| Rank | Model | Vendor | AIME | GPQA | Type | |---|---|---|---|---| | 🥇 | o3 | OpenAI | 96.7% | 89.2% | Closed | | 🥈 | Claude 4 Opus | Anthropic | 94.2% | 87.8% | Closed | | 🥉 | GPT-5.5 | OpenAI | 93.5% | 86.5% | Closed | | 4 | Gemini 3.1 | Google | 91.0% | 85.1% | Closed | | 5 | DeepSeek-V4 | DeepSeek | 88.3% | 82.0% | Open | | 6 | Claude 4 Sonnet | Anthropic | 86.0% | 80.5% | Closed | | 7 | GPT-5 | OpenAI | 84.5% | 79.0% | Closed | | 8 | ERNIE 5.1 | Baidu | 82.0% | 77.5% | Closed | | 9 | Qwen3-Max | Alibaba | 80.5% | 76.0% | Closed | | 10 | Gemini 3.0 | Google | 78.8% | 74.5% | Closed | | 11 | Kimi-2 | Moonshot | 77.0% | 73.0% | Closed | | 12 | Llama 4 Maverick | Meta | 75.5% | 71.5% | Open | | 13 | GLM-5 | Zhipu AI | 74.0% | 70.0% | Closed | | 14 | Mistral Large 3 | Mistral | 72.5% | 68.5% | Closed | | 15 | Claude 4 Haiku | Anthropic | 71.0% | 67.0% | Closed | | 16 | DeepSeek-V3.2 | DeepSeek | 69.5% | 65.5% | Open | | 17 | Llama 4 Scout | Meta | 68.0% | 64.0% | Open | | 18 | Yi-3 | 01.AI | 66.5% | 62.5% | Open | | 19 | Command A | Cohere | 65.0% | 61.0% | Closed | | 20 | MiniMax-M2.5 | MiniMax | 63.5% | 59.5% | Closed |
Model