Mathematical Reasoning on AIME 2025 (Pass@1 avg@64)
83.2Pass@1 Accuracy (avg@64)Klear-Reasoner-8B
Evaluation Results
| Method | Links | |
|---|---|---|
| Klear-Reasoner-8BMaximum Inference Length=64K2025.08 | 83.2 | |
| POLARIS-4B-PreviewMaximum Inference Length=96K2025.08 | 79.4 | |
| OpenReasoning-Nemotron-7BMaximum Inference Length=64K2025.08 | 78.2 | |
| Deepseek-R1-0528-Distill-8BMaximum Inference Length=64K2025.08 | 76.3 | |
| Klear-Reasoner-8BMaximum Inference Length=32K2025.08 | 75.6 | |
| Klear-Reasoner-8B-SFTMaximum Inference Length=32K2025.08 | 70.1 | |
| Qwen3-8BMaximum Inference Length=32K2025.08 | 67.3 | |
| AceReason-Nemotron-1.1-7BMaximum Inference Length=32K2025.08 | 64.8 | |
| MiMo-7B-RLMaximum Inference Length=32K2025.08 | 55.4 | |
| Skywork-OR1-7BMaximum Inference Length=32K2025.08 | 54.6 | |
| AReal-boba-RL-7BMaximum Inference Length=32K2025.08 | 48.3 |