diff --git a/README.md b/README.md index b270767..41c2cca 100644 --- a/README.md +++ b/README.md @@ -27,7 +27,7 @@ 1. [💡 Introduction](#-introduction) 2. [🎯 Current Results](#-current-results) -3. [🛠️ Installation](#%EF%B8%8F-installation) +3. [🛠️ Installation](#-installation) 4. [🚀 Usage](#-usage) 5. [🤝 Acknowledgements](#-acknowledgements) 6. [💬 Citation](#-citation) @@ -50,11 +50,35 @@ Whether you're interested in AI, automation, or contributing to cutting-edge age Results of Agent S2's Successful Rate (%) on the OSWorld full test set of all 369 test examples using Image input.
-| Benchmark | Agent S2 | Previous SOTA | Δ improve | -|-----------|----------|--------------|-----------| -| OSWorld (15 step) | 27.0% | 22.7% (ByteDance UI-TARS) | +4.3% | -| OSWorld (50 step) | 34.5% | 32.6% (OpenAI CUA) | +1.9% | -| AndroidWorld | 60.0% | 46.8% (ByteDance UI-TARS) | +13.2% | +| Benchmark | +Agent S2 | +Previous SOTA | +Δ improve | +
|---|---|---|---|
| OSWorld (15 step) | +27.0% | +22.7% (ByteDance UI-TARS) | ++4.3% | +
| OSWorld (50 step) | +34.5% | +32.6% (OpenAI CUA) | ++1.9% | +
| AndroidWorld | +60.0% | +46.8% (ByteDance UI-TARS) | ++13.2% | +