Devin Fusion Scores 68 in Coding Agent Benchmark at 31% Lower Cost Than Claude Code

Artificial Analysis tested Cognition's Devin Fusion CLI, which uses a dual-model architecture, scoring 68 at 31% lower cost than Claude Code. It also published its Coding Agent Index covering 39 agent-model configurations.

2026-10-21 ~ 2026-10-21 · 2 related posts

• Devin Fusion's dual-model setup scores 68 on Coding Agent Index at 31% lower cost than Claude Code — ArtificialAnlys · 2026-10-21
• Artificial Analysis opens up its Coding Agent Index: 39 configs, 3 benchmarks, full methodology — ArtificialAnlys · 2026-10-21