FailBench - robofac: leaderboard
Metric: Balanced Accuracy (%). Source: arxiv.org. Saturation forecast: Estimated already saturated. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemma 4 31B (IT) | 95 |
| 2 | Gemini 3 Flash | 90 |
| 3 | Qwen 3 VL 8B (Thinking) | 90 |
| 4 | GPT-4o | 85 |
| 5 | Qwen 3 VL 4B (Thinking) | 83 |
Interactive version: theaggregate.ai/benchmark?slug=failbench-robofac · How It Works · Data refreshed daily, snapshot 2026-09-24.