Python reference-model generation on RTLLM v2
77.3Pass@1ChipMate-Python-9B
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ChipMate-Python-9BType=ChipMate-P, Size=9B2026.05 | 77.3 | 81 | |
| ChipMate-Python-4BType=ChipMate-P, Size=4B2026.05 | 75.3 | 80.1 | |
| DeepSeek R1Type=Foundation Models, Size=671B2026.05 | 49.6 | 57.8 | |
| GPT-5.5Type=Foundation Models, Size=-2026.05 | 48.5 | 55.4 | |
| DeepSeek V4Type=Foundation Models, Size=1.6T2026.05 | 44.4 | 54.3 | |
| CodeV-R1Type=Specialized Models, Size=7B2026.05 | 42.7 | 51.8 | |
| DeepSeek CoderType=Foundation Models, Size=236B2026.05 | 42.4 | 50 | |
| Qwen3.5-9BType=Base Models, Size=9B2026.05 | 40.1 | 51.8 | |
| Qwen3.5-4BType=Base Models, Size=4B2026.05 | 38.2 | 46.1 | |
| CodeV-R1 (distill)Type=Specialized Models, Size=7B2026.05 | 36 | 46.2 |