Coding on LiveCodeBench (Pass@1)
78.86Pass@1DeepSeek-R1-0528
Evaluation Results
| Method | Links | |
|---|---|---|
| DeepSeek-R1-0528Quantization=W4A8-FP82026.02 | 78.86 | |
| DeepSeek-R1-0528Quantization=FP8-Block-Wise2026.02 | 77.1 | |
| GPT-4o2026.03 | 48.8 | |
| Llama-3.1-70B-Inst.2026.03 | 34.4 | |
| Eurus-2-7B-PRIME-R-TAPtraining=R-TAP integrated2026.03 | 31.8 | |
| Eurus-2-7B-PRIME2026.03 | 27.5 | |
| RLOO2026.03 | 26.7 | |
| TIC-GRPOModel=Qwen3-8B2025.08 | 21 | |
| GSPOModel=Qwen3-8B2025.08 | 20 | |
| GRPOModel=Qwen3-8B2025.08 | 19.88 | |
| Eurus-2-7B-SFT2026.03 | 17.8 | |
| TIC-GRPOModel=Qwen3-1.7B2025.08 | 12.11 | |
| Qwen2.5-Math-7B-Inst.2026.03 | 11.3 | |
| GSPOModel=Qwen3-1.7B2025.08 | 11.23 | |
| GRPOModel=Qwen3-1.7B2025.08 | 10.67 |