Devin Fusion 搭配 Claude Opus 5.5 与 GPT-6 Luna 编码成绩达 68.1
Devin Fusion 用 Opus 5.5 带一个更便宜的 GPT-6 Luna 副手，编码分数追平 Claude Code 还便宜三成，双模型怎么配合值得一试。
Cognition 的 Devin Fusion CLI 采用主力加副手双模型架构，本次测试组合为 Claude Opus 5.5 (max) 与 GPT-6 Luna (high)，主力模型负责规划、委派和最终审查。该组合在 Artificial Analysis Coding Agent Index 上得 68.1 分，与 Claude Code 加 Claude Sonnet 5.5 (max) 的 68.4 分接近，高于 Claude Code 加 Claude Opus 5.5 (max) 的 66.0 分。成本方面每任务 $9.01，比后者的 $13.04 低约 31%，平均耗时 55 分钟对 65 分钟。分项基准上，DeepSWE v1.1 得 72.6% 对 68.4%，SWE-Atlas-QnA 得 69.6% 对 66.4%，Terminal-Bench 4.0 略低为 62.1% 对 63.1%。
Devin Fusion with Claude Opus 5.5 (max) and a GPT-6 Luna (high) sidekick scores 68 on the Coding Agent Index, 2 points above Claude Code with Opus 5.5 (max) at 31% lower cost per task
@cognition's Devin Fusion runs a frontier lead model alongside a lower-cost sidekick model. The two work in parallel with their own context and tools, with the lead keeping planning, delegation and final review. We benchmarked Devin Fusion CLI with Claude Opus 5.5 (max) + GPT-6 Luna (high).
It scores 68.1 on the Artificial Analysis Coding Agent Index, closely matched at the top with Claude Code with Claude Sonnet 5.5 (max) at 68.4 and Muse Code with Muse Wonder Full (max) at 68.0. Against Claude Code with Claude Opus 5.5 (max) at 66.0, it costs $9.01 vs. $13.04 per task and averages 55 vs. 65 minutes per task.
Across the three benchmarks, the Fusion configuration scores higher on DeepSWE v1.1 (72.6% vs. 68.4%) and SWE-Atlas-QnA (69.6% vs. 66.4%) and slightly lower on Terminal-Bench 4.0 (62.1% vs. 63.1%). Its results reflect both models working together, including their combined API cost. Costs are pay-per-token API costs, not subscription prices.
