SceneActBench — leaderboard
Metric: Overall Score (%). Source: github.com. 11 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Opus 4.6 (High) | 48.9 |
| 2 | GPT-5.4 (High) | 48.7 |
| 3 | GPT-5.4 (Medium) | 48.7 |
| 4 | Gemini 3.1 Pro (Preview) (High) | 45.4 |
| 5 | Kimi K2.6 (Thinking) | 41.2 |
| 6 | Claude Sonnet 5 (High) | 39.5 |
Interactive version: theaggregate.ai/benchmark?slug=sceneactbench · How It Works · Data refreshed daily, snapshot 2026-08-05.