OmniACT: leaderboard
OmniACT: Measures browser, desktop, mobile, or GUI agents operating in interactive environments.
Metric: Action Score. Source: huggingface.co. Source: benchmarklist.com. Saturation forecast: Estimated already saturated. 15 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4 | 11.6 | #438 |
| 2 | Gemini 1.0 Pro | 11.46 | #933 |
| 3 | GPT-3.5 Turbo (0613) | 7.89 | #1118 |
| 4 | Palmyra X (43B) | 2.94 | #1317 |
| 5 | vicuna-13B | 1.78 | #1469 |
| 6 | LLaMA-13B | 1.62 | #1516 |
| 7 | Vicuna-7B | 0.77 | #1542 |
| 8 | LLaMA-7B | 0.48 | #1564 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=omniact · How It Works · Data refreshed daily, snapshot 2026-10-20.