WHAT NEW IN X10.31 SAT

X 上的 AI 合集档案 · 22 个合集 · 61,986 条

deepseek-v4.1POST

DeepSeek V4.1-Flash-1017 官方称 DeepSWE v1.1 得 77.0,作者建议先跑三个真实任

开源权重、API 价格不变,作者提醒跑分不等于自己的负载,换模型前先用三个真实任务验证。

Alex G
@AlexxGed12 天前收录原推发布于 2026.10.18
Alex G @AlexxGed原推发布于 2026.10.18
在 X 上打开原推

The cheap-model gap keeps shrinking. DeepSeek's V4.1-Flash-1017 is open-weight, keeps the same API price, and per DeepSeek hits 77.0 on DeepSWE v1.1. My rule: test 3 of your real tasks before switching. Benchmarks aren't your workload. Would you switch?

x.com/alexxged/status/2111714677122093156采集于 2026.10.18 16:34