Reels · 1 of 53
1:02
Reel · 1:02
Claude Code, Codex, OpenCode, OpenClaw, Hermes, mini-swe-agent and Raven on MiniMax-M3: 7 agent harnesses measured on Terminal-Bench 2.0
Narrated: seven agent harnesses drive the same MiniMax-M3 on 40 Terminal-Bench 2.0 tasks. The matrix of what each one solved, the same task in all seven at real speed, the cost of each harness and the verdict: the simplest came out on top, but no difference reaches significance.
- Length
- 1:02
- Published
This reel sums up Claude Code, Codex, OpenCode, OpenClaw, Hermes, mini-swe-agent and Raven on MiniMax-M3: 7 agent harnesses measured on Terminal-Bench 2.0, where the method, the tables and what did not work are.