deepseek-v4.1POST
V4.1-Flash-1017 跑分拆解：DeepSWE 与 Automation-Bench 领先，Terminal-
逐项对比：DeepSWE 77.0（前代 74.2，Sol 73.0，Opus 5 74.0）、Terminal-Bench 4.0 42.3（前代 31.2，Sol 39.9，Opus 5 51.8）、Automation-Bench 60.5（前代 54.8，Sol 45.8，Opus 5 50.3）。
Roman
@khudonogov10 天前收录原推发布于 2026.10.17
Roman @khudonogov原推发布于 2026.10.17
Pretty impressive: DeepSeek V4.1 Flash-1017 pulls ahead of GPT-5.6 Sol and Opus 5 on some coding and agent benchmarks, with much lower token prices. Open weights, an MIT license, and a 1M-token context window. Here’s the breakdown: - DeepSWE: 77.0, up from 74.2 for the previous Flash. GPT-5.6 Sol scores 73.0; Opus 5 scores 74.0. - Terminal-Bench 4.0: 42.3, up from 31.2. Ahead of GPT-5.6 Sol at 39.9, but behind Opus 5 at 51.8. - Automation-Bench: 60.5, up from 54.8. GPT-5.6 Sol scores 45.8; Opus 5 scores 50.3.
这条是引用着另一条写的：x.com/deepseek_ai/status/2111375640717033777
x.com/khudonogov/status/2111387033251836057采集于 2026.10.18 16:45