FrontierCode v1.1：V4.1 Flash 52.6%，仅次于 Opus 5.5

Artificial Analysis @ArtificialAnlys原推发布于 2026.10.23

在 X 上打开原推

FrontierCode v1.1 replaces DeepSWE v1.1 in the Index. Cognition's benchmark has 100 software engineering tasks on real open-source repositories with a private test set, graded on whether a maintainer would merge the change. Claude Code with Claude Opus 5.5 (max) scores 56.0%, ahead of Codex with DeepSeek V4.1 Flash (max) at 52.6%, Codex with GPT-6.1 Sol (max) at 49.4% and Codex with GPT-6 Astra (max) at 49.3%. Below the top, scores are tightly packed: 15 of the 24 configurations on the Index score between 43% and 49%.