Code Generation on LiveCodeBench v6 (Pass@1)
56.3Pass@1MiMo-RL 7B-R-TAP
Evaluation Results
| Method | Links | |
|---|---|---|
| MiMo-RL 7B-R-TAPModel Backbone=7B2026.03 | 56.3 | |
| MiMo 7B-RLModel Backbone=7B2026.03 | 49.3 | |
| OpenAI o1-miniModel Backbone=o1-mini2026.03 | 46.8 | |
| QwQ-32B PreviewModel Backbone=QwQ-32B2026.03 | 39.1 | |
| PriFT-massBackbone=Qwen3-4B-Instruct2026.06 | 37.25 | |
| Claude 3.5 Sonnet-1022Model Backbone=Claude 3.5 Sonnet2026.03 | 37.2 | |
| PriFT-probBackbone=Qwen3-4B-Instruct2026.06 | 36.87 | |
| R1-Distill-Qwen-14BModel Backbone=Qwen-14B2026.03 | 31.9 | |
| GPT-4o 0513Model Backbone=GPT-4o2026.03 | 30.9 | |
| OriginalBackbone=Qwen3-4B-Instruct2026.06 | 28.34 | |
| ASFTBackbone=Qwen3-4B-Instruct2026.06 | 26.73 | |
| IDFTBackbone=Qwen3-4B-Instruct2026.06 | 24.74 | |
| R1-Distill-Qwen-7BModel Backbone=Qwen-7B2026.03 | 23.9 | |
| TALRBackbone=Qwen3-4B-Instruct2026.06 | 23.6 | |
| DFTBackbone=Qwen3-4B-Instruct2026.06 | 23.22 | |
| EAFTBackbone=Qwen3-4B-Instruct2026.06 | 23.13 | |
| SFTBackbone=Qwen3-4B-Instruct2026.06 | 21.52 |