AgentBench
Input—per 1M tokens
Output—per 1M tokens
Context—tokens
WeightsClosed
About
AgentBench is ranked #15 of 29 in LLM evaluation tools on Inferse. It runs on Linux.
Compared on LLM evaluation tools
- Free plan
- Yesgithub.com
- Deployment options
- self-hostedgithub.com
Company
- Founded
- 2023github.com · 28 Sept 2026
Best AgentBench alternatives
See all 12Where it ranks on Inferse
- Best LLM Evaluation Tools in 2026#15 of 29
Sources
- github.com/THUDM/AgentBench· checked 28 Sept 2026




