EgoBench - Dynamic Easy Mode: leaderboard
Metric: Joint Success Rate (%; every required tool call issued and the final database state correct, cooperative simulated user revealing one step per turn). Source: arxiv.org. Saturation forecast: Around May 2028. 8 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Gemini 3.1 Pro (Preview) | 20.19 |
| 2 | Qwen 3.5 397B A17B (Non-reasoning) | 18.37 |
| 3 | Qwen 3.6 Plus (Non-reasoning) | 17.03 |
| 4 | Seed 2.0 Pro (Non-reasoning) | 14.74 |
| 5 | Kimi K2.5 (Non-reasoning) | 14.16 |
| 6 | MiMo-V2-Omni | 6.99 |
| 7 | GLM-5V Turbo | 6.67 |
| 8 | Qwen 3 VL 235B A22B Instruct | 1.24 |
Interactive version: theaggregate.ai/benchmark?slug=egobench-dynamic-easy-mode · How It Works · Data refreshed daily, snapshot 2026-09-25.