RoboAbstention - Ambiguous Referent: leaderboard
The Ambiguous Referent category of RoboAbstention (2026, The Yes-Man Syndrome): 420 instructions whose target matches several objects in the scene, given with an image to a vision-language robot planner that should abstain or ask. The score is the abstention rate in percent; higher is better.
Metric: Abstention rate (%) on the 420 Ambiguous Referent instructions, image plus instruction to an embodied VLM planner, every instruction warranting abstention, act or abstain judged by GPT-5.4 mini (97.5 percent agreement with human labels), API default sampling; higher is better. Source: arxiv.org. 8 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Gemini 2.5 Flash | 6 | #463 | |
| 2 | Llama 4 Maverick | 3.8 | #1110 | |
| 3 | Claude Sonnet 4.6 | 2.4 | #173 | |
| 4 | Gemini Robotics-ER 1.6 | 1.9 | #170 | |
| 5 | GPT-5.4 Mini | High | 1.7 | #406 |
| 6 | Qwen 3.5 27B | 1.4 | #375 | |
| 7 | GPT-5.4 Nano | 0.7 | #651 | |
| 8 | GPT-5.4 | 0 | #135 |
Interactive version: theaggregate.ai/benchmark?slug=roboabstention-ambiguous-referent · How It Works · Data refreshed daily, snapshot 2026-11-21.