Mathematical Reasoning on OlympiadBench
65.81AccuracyF-1
Evaluation results
| Method | Links | ||
|---|---|---|---|
| F-1Model=GPT-5, Model Category=Proprietary2026.01 | 65.81 | — | |
| CoTModel=GPT-5, Model Category=Proprietary2026.01 | 63.12 | — | |
| F-1Model=Gemini 2.5 Pro, Model Category=Proprietary2026.01 | 59.7 | — | |
| F-1Model=Qwen3-30B, Model Category=Open Source2026.01 | 59.68 | — | |
| Zero-ShotModel=GPT-5, Model Category=Proprietary2026.01 | 58.77 | — | |
| CoTModel=Qwen3-30B, Model Category=Open Source2026.01 | 58.23 | — | |
| CoTModel=Gemini 2.5 Pro, Model Category=Proprietary2026.01 | 57.67 | — | |
| F-1Model=Qwen3-235B, Model Category=Open Source2026.01 | 52.75 | — | |
| CoTModel=DeepSeek-V3.1, Model Category=Open Source2026.01 | 52.42 | — | |
| Zero-ShotModel=Qwen3-30B, Model Category=Open Source2026.01 | 51.67 | — | |
| F-1Model=DeepSeek-V3.1, Model Category=Open Source2026.01 | 50.29 | — | |
| PoTModel=Qwen3-30B, Model Category=Open Source2026.01 | 48.22 | — | |
| CoTModel=Qwen3-235B, Model Category=Open Source2026.01 | 48.03 | — | |
| Eurus-2-7B-PRIME-R-TAPtraining=R-TAP integrated2026.03 | 47.4 | — | |
| Zero-ShotModel=Qwen3-235B, Model Category=Open Source2026.01 | 47.33 | — | |
| AZR-R-TAPBase Model Type=Coder2026.03 | 46.6 | — | |
| PoTModel=Gemini 2.5 Pro, Model Category=Proprietary2026.01 | 46.29 | — | |
| Zero-ShotModel=DeepSeek-V3.1, Model Category=Open Source2026.01 | 45.24 | — | |
| AZR-R-TAPBase Model Type=Base2026.03 | 44.4 | — | |
| PoTModel=GPT-5, Model Category=Proprietary2026.01 | 43.72 | — | |
| GPT-4o2026.03 | 43.3 | — | |
| Zero-ShotModel=Gemini 2.5 Pro, Model Category=Proprietary2026.01 | 41.97 | — | |
| Qwen2.5-Math-7B-Inst.2026.03 | 40.7 | — | |
| Eurus-2-7B-PRIME2026.03 | 40.3 | — | |
| PoTModel=Qwen3-235B, Model Category=Open Source2026.01 | 38.56 | — | |
| RLOO2026.03 | 35.4 | — | |
| PoTModel=DeepSeek-V3.1, Model Category=Open Source2026.01 | 33.99 | — | |
| Llama-3.1-70B-Inst.2026.03 | 30.5 | — | |
| Eurus-2-7B-SFT2026.03 | 29.8 | — | |
| GPT-4o-202405132024.12 | 25.9 | — | |
| GHG-TDABase Model=Claude 3.5 Sonnet2026.02 | 19 | — | |
| GHG-TDABase Model=GPT-4o2026.02 | 18.3 | — | |
| GPT-4V2024.12 | 18 | — | |
| GoTBase Model=Claude 3.5 Sonnet2026.02 | 16 | — | |
| AoTBase Model=Claude 3.5 Sonnet2026.02 | 15.7 | — | |
| GoTBase Model=GPT-4o2026.02 | 15.3 | — | |
| AoTBase Model=GPT-4o2026.02 | 15 | — | |
| ToTBase Model=Claude 3.5 Sonnet2026.02 | 14.8 | — | |
| ToTBase Model=GPT-4o2026.02 | 14.1 | — | |
| CoTBase Model=Claude 3.5 Sonnet2026.02 | 12.9 | — | |
| CoTBase Model=GPT-4o2026.02 | 12.5 | — | |
| InternVL2.5-38Bparameters=38B2024.12 | 12.1 | — | |
| InternVL2.5-78Bparameters=78B2024.12 | 11.6 | — | |
| Qwen2-VL-72Bparameters=72B2024.12 | 11.2 | — | |
| InternVL2.5-26Bparameters=26B2024.12 | 8.8 | — | |
| InternVL2-Llama3-76Bparameters=76B, backbone=Llama32024.12 | 5.5 | — | |
| InternVL2.5-8Bparameters=8B2024.12 | 4.9 | — | |
| InternVL2-40Bparameters=40B2024.12 | 3.9 | — | |
| InternVL2-26Bparameters=26B2024.12 | 3.5 | — | |
| InternVL2.5-4Bparameters=4B2024.12 | 3 | — | |
| InternVL2.5-2Bparameters=2B2024.12 | 2 | — | |
| InternVL2-8Bparameters=8B2024.12 | 1.9 | — | |
| InternVL2.5-1Bparameters=1B2024.12 | 1.7 | — | |
| InternVL2-4Bparameters=4B2024.12 | 1.1 | — | |
| InternVL-Chat-V1.52024.12 | 0.6 | — | |
| InternVL2-2Bparameters=2B2024.12 | 0.4 | — | |
| InternVL2-1Bparameters=1B2024.12 | 0.3 | — | |
| AoTBackbone=Llama 3-8B2026.02 | — | 10.89 | |
| AoTBackbone=Qwen 2-14B2026.02 | — | 7.8 | |
| CoTBackbone=Llama 3-8B2026.02 | — | 8.37 | |
| CoTBackbone=Qwen 2-14B2026.02 | — | 6.4 | |
| GHS-TDABackbone=Llama 3-8B2026.02 | — | 13.05 | |
| GHS-TDABackbone=Qwen 2-14B2026.02 | — | 9.7 | |
| GoTBackbone=Llama 3-8B2026.02 | — | 11.79 | |
| GoTBackbone=Qwen 2-14B2026.02 | — | 8.4 | |
| ToTBackbone=Llama 3-8B2026.02 | — | 10.26 | |
| ToTBackbone=Qwen 2-14B2026.02 | — | 7.2 |