RoboStressBench - State Understanding: leaderboard
Metric: Accuracy (%) on the 633 object-state-understanding multiple-choice questions, exact match on the predicted option, greedy decoding with at most 64 new tokens (API models: temperature 0, 64 output tokens), one unified instruction template; images of embodied scenes under physical visual stress (material, viewpoint, lighting, geometry); higher is better. Source: arxiv.org. Saturation forecast: Around 2030. 16 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3.6 27B | 67.1 |
| 2 | Qwen 3.6 35B A3B | 67.1 |
| 3 | Qwen 3.5 27B | 63.3 |
| 4 | Qwen 3.5 35B A3B | 62.9 |
| 5 | Qwen 3.5 4B | 59.6 |
| 6 | Qwen 3.5 9B | 58.9 |
| 7 | Gemini 3.1 Pro (Preview) | 41.6 |
| 8 | InternVL3.5-8B | 41.5 |
| 9 | Molmo2-8B | 39.2 |
| 10 | GPT-5.5 | 38.9 |
Interactive version: theaggregate.ai/benchmark?slug=robostressbench-state-understanding · How It Works · Data refreshed daily, snapshot 2026-09-29.