← Benchmarks
AIME 2025
All 30 problems from the 2025 American Invitational Mathematics Examination (AIME I and AIME II), testing olympiad-level mathematical reasoning with integer answers from 000-999. Used as an AI benchmark to evaluate large language models' ability to solve complex mathematical problems requiring multi-step logical deductions and structured symbolic reasoning.
id aime-2025 · max 1 · 122 models reported
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3 Pro Google | 1.00 |
| 2 | GPT-5.2 OpenAI | 1.00 |
| 3 | GPT-5.2 Pro OpenAI | 1.00 |
| 4 | Grok-4 Heavy xAI | 1.00 |
| 5 | Kimi K2-Thinking-0905 Moonshot AI · open | 1.00 |
| 6 | Claude Opus 4.6 Anthropic | 1.00 |
| 7 | Gemini 3 Flash Google | 1.00 |
| 8 | GPT-5.1 High OpenAI | 1.00 |
| 9 | LongCat-Flash-Thinking-2601 Meituan · open | 1.00 |
| 10 | Nemotron 3 Nano (30B A3B) NVIDIA · open | 0.99 |
| 11 | GPT OSS 20B High OpenAI · open | 0.99 |
| 12 | GPT-5.1 Medium OpenAI | 0.98 |
| 13 | Seed 2.0 Pro ByteDance | 0.98 |
| 14 | Step-3.5-Flash StepFun · open | 0.97 |
| 15 | MAI-Thinking-1 Microsoft | 0.97 |
| 16 | GPT-5.1 Codex High OpenAI | 0.97 |
| 17 | Sarvam-105B Sarvam AI · open | 0.97 |
| 18 | Sarvam-30B Sarvam AI · open | 0.97 |
| 19 | Kimi K2.5 Moonshot AI · open | 0.96 |
| 20 | DeepSeek-V3.2-Speciale DeepSeek · open | 0.96 |
| 21 | GLM-4.7 Zhipu AI · open | 0.96 |
| 22 | GPT-5 OpenAI | 0.95 |
| 23 | GPT-5 High OpenAI | 0.95 |
| 24 | Seed 1.8 ByteDance | 0.94 |
| 25 | MiMo-V2-Flash Xiaomi · open | 0.94 |
| 26 | GPT-5.1 OpenAI | 0.94 |
| 27 | GPT-5.1 Instant OpenAI | 0.94 |
| 28 | GPT-5.1 Thinking OpenAI | 0.94 |
| 29 | GLM-4.6 Zhipu AI · open | 0.94 |
| 30 | Grok-3 xAI | 0.93 |
| 31 | DeepSeek-V3.2 DeepSeek · open | 0.93 |
| 32 | DeepSeek-V3.2 (Thinking) DeepSeek · open | 0.93 |
| 33 | Seed 2.0 Lite ByteDance | 0.93 |
| 34 | EXAONE 4.5 33B LG AI Research | 0.93 |
| 35 | K-EXAONE-236B-A23B LG AI Research | 0.93 |
| 36 | o4-mini OpenAI | 0.93 |
| 37 | GPT OSS 120B High OpenAI · open | 0.93 |
| 38 | Nova 2 Pro Amazon | 0.92 |
| 39 | Qwen3-235B-A22B-Thinking-2507 Alibaba Cloud / Qwen Team · open | 0.92 |
| 40 | Nova 2 Omni Amazon | 0.92 |
| 41 | Grok 4 Fast xAI | 0.92 |
| 42 | Grok-4 xAI | 0.92 |
| 43 | GLM-4.7-Flash Zhipu AI · open | 0.92 |
| 44 | GPT-5 mini OpenAI | 0.91 |
| 45 | Mercury 2 Inception | 0.91 |
| 46 | Nova 2 Lite Amazon | 0.91 |
| 47 | Grok-3 Mini xAI | 0.91 |
| 48 | LongCat-Flash-Thinking Meituan · open | 0.91 |
| 49 | Nemotron 3 Super (120B A12B) NVIDIA · open | 0.90 |
| 50 | Command A+ Cohere · open | 0.90 |
| 51 | Qwen3 VL 235B A22B Thinking Alibaba Cloud / Qwen Team · open | 0.90 |
| 52 | DeepSeek-V3.2-Exp DeepSeek · open | 0.89 |
| 53 | IBM Granite 4.2 30B IBM · open | 0.89 |
| 54 | GPT-5 Medium OpenAI | 0.89 |
| 55 | Gemini 2.5 Pro Preview 06-05 Google | 0.88 |
| 56 | Qwen3-Next-80B-A3B-Thinking Alibaba Cloud / Qwen Team · open | 0.88 |
| 57 | Step3-VL-10B StepFun · open | 0.88 |
| 58 | DeepSeek-R1-0528 DeepSeek · open | 0.88 |
| 59 | Claude Sonnet 4.5 Anthropic | 0.87 |
| 60 | ERNIE 5.0 Baidu | 0.87 |
| 61 | Seed 2.0 Mini ByteDance | 0.87 |
| 62 | IBM Granite 4.2 8B IBM · open | 0.87 |
| 63 | o3 OpenAI | 0.86 |
| 64 | Mistral Medium 3.5 Mistral AI · open | 0.86 |
| 65 | GPT-5 nano OpenAI | 0.85 |
| 66 | Ministral 3 (14B Reasoning 2512) Mistral AI · open | 0.85 |
| 67 | Mistral Small 4 Mistral AI · open | 0.84 |
| 68 | Qwen3 VL 32B Thinking Alibaba Cloud / Qwen Team · open | 0.84 |
| 69 | Qwen3 VL 30B A3B Thinking Alibaba Cloud / Qwen Team · open | 0.83 |
| 70 | Gemini 2.5 Pro Google | 0.83 |
| 71 | Qwen3 Max Alibaba Cloud / Qwen Team | 0.82 |
| 72 | Qwen3 235B A22B Alibaba Cloud / Qwen Team · open | 0.81 |
| 73 | GPT-5.5 Instant OpenAI | 0.81 |
| 74 | MiniMax M2.1 MiniMax · open | 0.81 |
| 75 | Claude Haiku 4.5 Anthropic | 0.81 |
| 76 | Qwen3 VL 8B Thinking Alibaba Cloud / Qwen Team · open | 0.80 |
| 77 | Ministral 3 (8B Reasoning 2512) Mistral AI · open | 0.79 |
| 78 | IBM Granite 4.2 3B IBM · open | 0.78 |
| 79 | MiniCPM-SALA OpenBMB · open | 0.78 |
| 80 | Claude Opus 4.1 Anthropic | 0.78 |
| 81 | MiniMax M2 MiniMax · open | 0.78 |
| 82 | Phi 4 Reasoning Plus Microsoft · open | 0.78 |
| 83 | MiniMax M1 80K MiniMax · open | 0.77 |
| 84 | Claude Opus 4 Anthropic | 0.76 |
| 85 | Qwen3 VL 235B A22B Instruct Alibaba Cloud / Qwen Team · open | 0.75 |
| 86 | MiniMax M1 40K MiniMax · open | 0.75 |
| 87 | Qwen3 VL 4B Thinking Alibaba Cloud / Qwen Team · open | 0.74 |
| 88 | Qwen3 32B Alibaba Cloud / Qwen Team · open | 0.73 |
| 89 | Llama 3.1 Nemotron Ultra 253B v1 NVIDIA · open | 0.72 |
| 90 | Min istral 3 (3B Reasoning 2512) Mistral AI · open | 0.72 |
| 91 | Nemotron Nano 9B v2 NVIDIA · open | 0.72 |
| 92 | Gemini 2.5 Flash Google | 0.72 |
| 93 | Qwen3 30B A3B Alibaba Cloud / Qwen Team · open | 0.71 |
| 94 | Claude Sonnet 4 Anthropic | 0.70 |
| 95 | Qwen3 14B Alibaba Cloud / Qwen Team · open | 0.70 |
| 96 | Qwen3-235B-A22B-Instruct-2507 Alibaba Cloud / Qwen Team · open | 0.70 |
| 97 | Qwen3-Next-80B-A3B-Instruct Alibaba Cloud / Qwen Team · open | 0.69 |
| 98 | Qwen3 VL 30B A3B Instruct Alibaba Cloud / Qwen Team · open | 0.69 |
| 99 | Qwen3 VL 32B Instruct Alibaba Cloud / Qwen Team · open | 0.66 |
| 100 | Magistral Medium Mistral AI · open | 0.65 |