Trending storyDeveloping

Devin Fusion CLI scores 68.1 on Artificial Analysis Coding Agent Index

1 article1 sourcesince Oct 21Last article 1h ago · Oct 21, 1:10 AM ET

Overview

AISummary of 1 article

Artificial Analysis benchmarked Devin Fusion CLI, paired with Claude Opus 5.5 (max) and GPT-6 Luna (high), and it scored 68.1 on its Coding Agent Index.

Claude Code with Claude Opus 5.5 (max) scored 66.0 in the same index. Artificial Analysis reports Fusion cost $9.01 per task versus $13.04 for Claude Code, and took 55 minutes on average versus 65.

Fusion scored higher than Claude Code on DeepSWE v1.1 and SWE-Atlas-QnA, and slightly lower on Terminal-Bench 4.0. Artificial Analysis says the costs are pay-per-token API costs, not subscription prices. The index's top entries are close: Claude Code with Claude Sonnet 5.5 (max) scored 68.4 and Muse Code with Muse Wonder Full (max) scored 68.0.

Written by AI from the articles below · updated Oct 21, 1:24 AM ET

Check the sources:

Latest developmentOct 21, 1:10 AM ET

Devin Fusion scores 68.1 on Coding Agent Index at lower cost per task

Article timeline

The articles in this story. Times are ET.

Oct 21

1:10 AM
Artificial Analysis@ArtificialAnlysOfficialPick
Devin Fusion scores 68.1 on Coding Agent Index at lower cost per task
AIArtificial Analysis benchmarked Devin Fusion CLI, pairing Claude Opus 5.5 (max) with GPT-6 Luna (high), and it scored 68.1 on its Coding Agent Index. Against Claude Code with Claude Opus 5.5 (max) at 66.0, Fusion cost $9.01 versus $13.04 per task and took 55 versus 65 minutes on average. Fusion scored higher on DeepSWE v1.1 and SWE-Atlas-QnA but slightly lower on Terminal-Bench 4.0, and costs are pay-per-token API costs.

Heat Trend

Not enough continuous observations to show a trend yet.