RuleSafe-VL - Decision State Prediction: leaderboard
Metric: Macro-F1 (%; decidable versus underdetermined image-text cases, mean of three policy families). Source: arxiv.org. Saturation forecast: Around 2030. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Qwen 3 VL 4B | 64.5 |
| 2 | Qwen 3 VL 8B | 59.2 |
| 3 | Claude Opus 4.6 | 58.6 |
| 4 | GPT-5.4 | 51.7 |
| 5 | Gemini 3.1 Pro (Preview) | 38.6 |
| 6 | Qwen 2.5 VL 7B Instruct | 34.6 |
Interactive version: theaggregate.ai/benchmark?slug=rulesafe-vl-decision-state-prediction · How It Works · Data refreshed daily, snapshot 2026-09-25.