Robot Failure Detection on BridgeData Fail V2
85Execution AccuracyGuardian-8B-Thinking
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Guardian-8B-ThinkingModel Category=Specialized / small models, Reasoning Mode=Thinking (explicit reasoning)2025.12 | 85 | 91 | |
| Qwen3-VL-235B-A22BModel Category=Large-scale generalist models2025.12 | 75 | 86 | |
| GPT4.1Model Category=Large-scale generalist models2025.12 | 72 | 86 | |
| InternVL3-8BModel Category=Specialized / small models2025.12 | 66 | 71 | |
| CLIP+MLPModel Category=Specialized / small models2025.12 | 58 | 54 | |
| SentinelModel Category=Specialized / small models2025.12 | 57 | — | |
| Cosmos-Reason1-7BModel Category=Specialized / small models2025.12 | 53 | 61 |