Robotic Planning on LEMMA Single-Agent Multiplicity
77Success Rate (SR)SG-CoT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SG-CoTLLM=Gemini-2.5-Flash2026.03 | 77 | 51 | |
| SG-CoTLLM=Qwen3-VL-2B2026.03 | 71 | 52 | |
| InnerMonoLLM=Gemini-2.5-Flash2026.03 | 70 | 62 | |
| InnerMonoLLM=Qwen3-VL-2B2026.03 | 65 | 53 | |
| SG-CoTLLM=Gemini-2.5-Flash, Scene Graph Edges=false2026.03 | 59 | 32 | |
| SG-CoTLLM=Gemini-2.5-Flash, Iterations=12026.03 | 52 | 40 | |
| SG-CoTLLM=Qwen3-VL-2B, Scene Graph Edges=false2026.03 | 52 | 37 | |
| CLARALLM=Gemini-2.5-Flash2026.03 | 48 | 47 | |
| SG-CoTLLM=Qwen3-VL-2B, Iterations=12026.03 | 45 | 35 | |
| CLARALLM=Qwen3-VL-2B2026.03 | 38 | 35 | |
| SG-CoTLLM=Qwen3-VL-2B, Iterations=02026.03 | 34 | 32 | |
| SG-CoTLLM=Gemini-2.5-Flash, Iterations=02026.03 | 32 | 28 | |
| ProgPromptLLM=Gemini-2.5-Flash2026.03 | 26 | 24 | |
| ProgPromptLLM=Qwen3-VL-2B2026.03 | 25 | 23 |