modelbenchmark.io

Agent Prime

agent-prime

Compare

Reliable models for dependable agentic applications, multi-step tool use, and reasoning workflows. Any model that meets the contract spec can serve your request.

Specification

most-agreed values

Context
197K
Max output
30K
Released
2026-05-04
Knowledge cutoff
Retires
Open weights
no
Input
text
Output
text

Available from 1 host

HostIn $/MOut $/MCache rdCache wrContextOutputRetires
the-grid-ai197K30K