Planning Accuracy on Blocksworld (test)
98.2AccuracyDeepSeek-R1
Evaluation Results
| Method | Links | |
|---|---|---|
| DeepSeek-R12026.07 | 98.2 | |
| o1-preview2026.07 | 97.8 | |
| PT-SFTBase model=Mistral2026.05 | 97 | |
| PT-SFTBase model=Qwen2.52026.05 | 96 | |
| PT-SFTBase model=GPT-OSS2026.05 | 93 | |
| QwQ-32B2026.07 | 88.8 | |
| HyperGuideBase model=Qwen2.52026.05 | 87 | |
| HyperGuideBase model=GPT-OSS2026.05 | 83 | |
| SoftCoTBase model=Qwen2.52026.05 | 82 | |
| OVMBase model=Qwen2.52026.05 | 81.4 | |
| SoftCoTBase model=GPT-OSS2026.05 | 79 | |
| OVMBase model=GPT-OSS2026.05 | 77.4 | |
| Qwen3-8BTraining=CARL2026.07 | 77.2 | |
| HyperGuideBase model=Mistral2026.05 | 76 | |
| Qwen3-8BTraining=RFT2026.07 | 73.8 | |
| SoftCoTBase model=Mistral2026.05 | 71 | |
| OVMBase model=Mistral2026.05 | 62 | |
| Self-ConsistencyBase model=Qwen2.52026.05 | 60 | |
| Tree of ThoughtsBase model=Qwen2.52026.05 | 58 | |
| Few-shotBase model=Mistral2026.05 | 57 | |
| Tree of ThoughtsBase model=Mistral2026.05 | 55 | |
| Self-ConsistencyBase model=Mistral2026.05 | 53 | |
| DeepSeek-R1-Distill-Llama-8BTraining=CARL2026.07 | 52.6 | |
| DeepSeek-V32026.07 | 44.8 | |
| GPT-4o2026.07 | 42.4 | |
| DeepSeek-R1-Distill-Llama-8BTraining=RFT2026.07 | 42 | |
| Few-shotBase model=Qwen2.52026.05 | 41 | |
| Tree of ThoughtsBase model=GPT-OSS2026.05 | 37 | |
| Qwen3-8B2026.07 | 31.2 | |
| Llama-3.1-70B-Instruct2026.07 | 21.6 | |
| Qwen2.5-72B-Instruct2026.07 | 13.8 | |
| Few-shotBase model=GPT-OSS2026.05 | 9 | |
| Self-ConsistencyBase model=GPT-OSS2026.05 | 2 | |
| DeepSeek-R1-Distill-Llama-8B2026.07 | 1.4 | |
| Llama-3.1-8B-Instruct2026.07 | 0.6 |