Agent Arena - Confirmed Success — leaderboard
Metric: Confirmed Success (%). Source: arena.ai. 38 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Nemotron 3 Ultra | 15.08 |
| 2 | Kimi K3 | 14 |
| 3 | Gemini 3.5 Flash (Medium) | 13.18 |
| 4 | Claude Fable 5 (High) | 10.67 |
| 5 | Claude Opus 4.8 (Thinking) | 8.9 |
| 6 | Gemini 3 Flash | 8.74 |
| 7 | GLM-5.2 (Max) | 8.65 |
| 8 | Claude Sonnet 5 (High) | 8.14 |
| 9 | GPT-5.6 Sol (xHigh) | 7.25 |
| 10 | Claude Opus 4.8 | 7.1 |
| 11 | GPT-5.5 (xHigh) | 6.65 |
| 12 | GPT-5.4 (High) | 6.23 |
| 13 | GPT-5.5 (High) | 6.2 |
| 14 | Claude Opus 4.7 (Thinking) | 5.67 |
| 15 | Claude Opus 4.7 | 4.97 |
Interactive version: theaggregate.ai/benchmark?slug=agent-arena-confirmed-success · How the rankings work · Data refreshed daily, snapshot 2026-07-22.