List your company

Agentic Tool Calling on MAVEN-Bench

71AccuracyMAVEN
30.4440.9751.562.03May 29, 2026

Evaluation results

MethodLinks
MAVENBackbone=GPT-OSS-120b, Price=$1.52026.057195.167.418.29.5
Claude-Sonnet-4.5Price=$152026.057094.26717.79.5
Kimi-K2Price=$52026.055792.465.517.79.3
Grok-4Price=$152026.055588.26316.88.4
GPT-OSS-120bPrice=$0.92026.054888.962.617.58.7
GLM-4.5Price=$32026.05438862.316.98.7
o4-miniPrice=$92026.053876.753.615.77.5
GPT-5Price=$152026.053261.143.312.65.2