deepseek-v4.1POST

FrontierCode v1.1: DeepSeek V4.1 Flash 1017 beats GPT-6.1 Sol

On Cognition's FrontierCode v1.1 benchmark of 100 real open-source software engineering tasks with a private test set, DeepSeek V4.1 Flash 1017 MAX ranks above GPT-6.1 Sol MAX.

Leo · Building w/ AI

@leo114119Archived 9 days agoOriginally posted 2026.10.23

Leo · Building w/ AI @leo114119Originally posted 2026.10.23

Open original on X

FrontierCode v1.1 在 Index 中取代了 DeepSWE v1.1。Cognition 的基准测试包含 100 个软件工程任务，这些任务基于真实的开源仓库，并使用私有测试集进行评估 Deepseek v4.1 Flash 1017 MAX 居然超过了 GPT 6.1 SOL MAX 太疯狂了吧。。。这意味着在真实的编码任务里Deepseek v4.1 Flash 1017比 GPT 6.1 SOL更好用

This post quotes another post: “FrontierCode v1.1: DeepSeek V4.1 Flash scores 52.6%, behind Opus 5.5”

x.com/leo114119/status/2113479417003459022Archived 2026.10.23 12:06