Ko-AgentBench - L4 Parallel Tool Reasoning — leaderboard

Metric: Coverage (%). Source: huggingface.co. 13 models tracked.

Top models

#ModelScore
1Claude Sonnet 4.575
2Claude Haiku 4.575
3GPT-4o71.67
4Qwen 3 Next 80B A3B70
5DeepSeek V3.166.67
6nova-2-lite-v165
7GPT-4o Mini58.33
8Grok 4.1 Fast56.67
9GPT-555
10Gemini 2.5 Pro46.67
11GPT-5 Mini36.67
12Gemini 2.5 Flash35
13Gemini 2.5 Flash Lite35

Interactive version: aibenchmarks.dev/benchmark?slug=ko-agentbench-l4-parallel-tool-reasoning · How the rankings work · Data refreshed daily, snapshot 2026-07-17.