Embodied Task Planning on Robotouille Synchronous
97Pass@1 AccuracyGiG+Exp
Evaluation Results
| Method | Links | |
|---|---|---|
| GiG+ExpLLM Backbone=Qwen3-235B-A22B-Instruct-2507-FP82026.01 | 97 | |
| GiGLLM Backbone=Qwen3-235B-A22B-Instruct-2507-FP82026.01 | 93 | |
| GiGLLM Backbone=Gemini-2.5-Flash2026.01 | 92 | |
| ReActLLM Backbone=Gemini-2.5-Flash2026.01 | 92 | |
| GiGLLM Backbone=DeepSeek-R1-FP42026.01 | 91 | |
| GiG+ExpLLM Backbone=Gemini-2.5-Flash2026.01 | 90 | |
| ReCAPLLM Backbone=Gemini-2.5-Flash2026.01 | 89 | |
| GiG+ExpLLM Backbone=DeepSeek-R1-FP42026.01 | 88 | |
| ReActLLM Backbone=Qwen3-235B-A22B-Instruct-2507-FP82026.01 | 74 | |
| ReCAPLLM Backbone=DeepSeek-R1-FP42026.01 | 72 | |
| ReCAPLLM Backbone=Qwen3-235B-A22B-Instruct-2507-FP82026.01 | 71 | |
| ReActLLM Backbone=DeepSeek-R1-FP42026.01 | 53 | |
| CoTLLM Backbone=Gemini-2.5-Flash2026.01 | 34 | |
| CoTLLM Backbone=Qwen3-235B-A22B-Instruct-2507-FP82026.01 | 7 | |
| CoTLLM Backbone=DeepSeek-R1-FP42026.01 | 2 |