Mistral Vibe for Code scored 22/25 in a capability comparison of coding agents, ahead of Claude Code and OpenAI Codex, which each scored 21/25. The evaluation tested Mistral, Claude, Cursor, and OpenAI across scaffolding, testing, and pull request workflows. Mistral led on cost and surface coverage, while Claude Code led on raw execution. The comparison relied on vendor benchmarks, which the author notes are not directly comparable across different suites.
No score is assigned. Sources and their independence are shown in the citation chain below.