List your company

Robotic Planning on LEMMA Stack and Pass tasks (test)

75Success RateSG-CoT
37.5647.285766.72Mar 18, 2026

Evaluation results

MethodLinks
SG-CoTLLM Backbone=Gemini-2.5-Flash2026.0375
ProgPromptLLM Backbone=Gemini-2.5-Flash2026.0359
SG-CoTLLM Backbone=Qwen3-VL-2B2026.0359
InnerMonoLLM Backbone=Gemini-2.5-Flash2026.0348
InnerMonoLLM Backbone=Qwen3-VL-2B2026.0345
CLARALLM Backbone=Gemini-2.5-Flash2026.0345
ProgPromptLLM Backbone=Qwen3-VL-2B2026.0344
CLARALLM Backbone=Qwen3-VL-2B2026.0339