Planning on PlanBench
99.1AccuracyDeepSeek-R1 Distill
Evaluation Results
| Method | Links | |
|---|---|---|
| DeepSeek-R1 DistillSize=70B2026.06 | 99.1 | |
| TeamTR (Heterogeneous)Size=1.7B+8B+14B2026.05 | 44.5 | |
| TeamTR (Homogeneous)Size=3×8B2026.05 | 44.1 | |
| Qwen3-Base (DICE-PC)Size=3×4B2026.06 | 42.7 | |
| Qwen3-235B-A22BSize=235B2026.06 | 42.5 | |
| OpenAI o3Size=–2026.06 | 42.1 | |
| Qwen3-Base (DICE-FT)Size=3×4B2026.06 | 41.9 | |
| o4-miniSize=–2026.05 | 41.2 | |
| OpenAI o3-miniSize=–2026.06 | 41.2 | |
| Qwen3 (thinking)Size=32B2026.05 | 40.3 | |
| Qwen3 (thinking)Size=32B2026.06 | 40.3 | |
| Qwen3-Inst (Few-shot CoT)Size=4B2026.06 | 40.1 | |
| Qwen3-Inst (SC@64)Size=14B2026.06 | 39.7 | |
| Qwen3-A3BSize=30B2026.05 | 39.1 | |
| DeepSeek-R1 DistillSize=70B2026.05 | 39.1 | |
| Debate + JudgeSize=3×8B2026.05 | 39.1 | |
| Qwen3-Base Debate+Judge (FT)Size=3×8B2026.06 | 39.1 | |
| Qwen3-A3BSize=30B2026.06 | 39.1 | |
| Role-playSize=3×8B2026.05 | 38.7 | |
| Qwen3-Base Role-play (FT)Size=3×8B2026.06 | 38.7 | |
| Qwen3-Base Debate+Judge (FT)Size=3×4B2026.06 | 38.4 | |
| DebateSize=3×8B2026.05 | 38.3 | |
| Qwen3-Base Debate (FT)Size=3×8B2026.06 | 38.3 | |
| Qwen3-Inst Debate+Judge (prompt)Size=3×4B2026.06 | 38.1 | |
| QwQSize=32B2026.05 | 37.9 | |
| Qwen3-Inst (ReAct)Size=4B2026.06 | 37.9 | |
| Qwen3-Base Role-play (FT)Size=3×4B2026.06 | 37.9 | |
| QwQSize=32B2026.06 | 37.9 | |
| Qwen3-Inst Role-play (prompt)Size=3×4B2026.06 | 37.8 | |
| Qwen3-Base Debate (FT)Size=3×4B2026.06 | 37.5 | |
| Qwen3-Inst Debate (prompt)Size=3×4B2026.06 | 37.2 | |
| Llama 3.1-Base (DICE-PC)Size=3×8B2026.06 | 37.1 | |
| Qwen3-Inst (ToT)Size=4B2026.06 | 37 | |
| Qwen3-Inst (SC@64)Size=4B2026.06 | 36.3 | |
| Qwen3-Base-DAPOSize=14B2026.06 | 35.6 | |
| Qwen3-Base-GRPOSize=14B2026.06 | 34.8 | |
| GPT-4o-miniSize=–2026.05 | 34.7 | |
| GPT-4o-miniSize=–2026.06 | 34.7 | |
| Qwen3-BaseSize=32B2026.06 | 34.3 | |
| Qwen2.5-BaseSize=72B2026.06 | 33.7 | |
| DAPOSize=8B2026.05 | 33.5 | |
| Qwen3-Base-DAPOSize=8B2026.06 | 33.5 | |
| GRPOSize=8B2026.05 | 33.1 | |
| Qwen3-Base-GRPOSize=8B2026.06 | 33.1 | |
| Qwen3-Base-PPOSize=8B2026.06 | 32.6 | |
| Qwen3-BaseSize=14B2026.06 | 29.7 | |
| Qwen3-Inst (Zero-shot CoT)Size=4B2026.06 | 29.5 |