deepseek-v4.1POST
DeepSeek V4.1 Flash 1017 发布：DeepSWE 77.0 略胜 Opus 5
开源权重 MIT，DeepSWE 77.0 对 Opus 5 的 74.0，每任务约 $0.20 便宜约 50 倍；但 Terminal-Bench 与 ProgramBench 仍落后。
Oguzhan
@0xOgi11 天前收录原推发布于 2026.10.17
Oguzhan @0xOgi原推发布于 2026.10.17
77.0 on DeepSWE. DeepSeek-V4.1-Flash-1017 is out today, open weights, MIT. On DeepSeek's chart it edges Opus 5 (74.0) on DeepSWE, about $0.20 a task, roughly 50x cheaper. Opus 5 still leads Terminal-Bench 51.8 to 42.3 and ProgramBench 37.0 to 29.1. Send routine fixes to Flash.
x.com/0xogi/status/2111479884157419832采集于 2026.10.18 16:35