Ko-AgentBench - L5 Error Handling & Robustness — leaderboard
Metric: Adaptive Routing Score (%). Source: huggingface.co. 13 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Grok 4.1 Fast | 34.75 |
| 2 | GPT-5 | 30 |
| 3 | nova-2-lite-v1 | 26.25 |
| 4 | GPT-4o | 25.83 |
| 5 | Qwen 3 Next 80B A3B | 25.42 |
| 6 | Claude Haiku 4.5 | 22.83 |
| 7 | GPT-4o Mini | 21.71 |
| 8 | DeepSeek V3.1 | 21.33 |
| 9 | Claude Sonnet 4.5 | 18.92 |
| 10 | Gemini 2.5 Pro | 12.5 |
| 11 | Gemini 2.5 Flash Lite | 12.5 |
| 12 | Gemini 2.5 Flash | 10 |
| 13 | GPT-5 Mini | 9.17 |
Interactive version: aibenchmarks.dev/benchmark?slug=ko-agentbench-l5-error-handling-robustness · How the rankings work · Data refreshed daily, snapshot 2026-07-17.