WildAgtEval: leaderboard
Agent API-use benchmark measuring robustness to real-world API complexity across multi-turn conversations, API functions, and injected complexity scenarios.
Metric: Complexity-Injected API Call Accuracy (self-reported). Source: benchmarklist.com. Status: saturation imminent. 10 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Claude Sonnet 4 (Thinking) | 67.5 |
| 2 | Claude Sonnet 4 | 63.6 |
| 3 | Qwen 3 235B A22B (Thinking) | 62.6 |
| 4 | GPT-OSS-120B | 62.5 |
| 5 | Claude 3.7 Sonnet | 61.6 |
| 6 | Claude 3.5 Sonnet | 55.8 |
| 7 | Qwen 3 32B | 49.9 |
Interactive version: theaggregate.ai/benchmark?slug=wildagteval · How It Works · Data refreshed daily, snapshot 2026-09-05.