DeepSeek-V4.1-Flash-1017 on AA & Vals index
英文标题：DeepSeek-V4.1-Flash-1017 on AA & Vals index
DeepSeek-V4.1-Flash-1017's Terminal bench 4.0 score: - Model card reported 42.3% - AA reports 34% - Vals index reports 31.82% AA Vals Index Model card Model card quotes: We use the Claude Code harness for Terminal-Bench 4.0 I am guessing AA & Vals index uses mini-swe-agent thus causing a low
打开原文相关 10 · 机会 1
中文
DeepSeek-V4.1-Flash-1017's Terminal bench 4.0 score: - Model card reported 42.3% - AA reports 34% - Vals index reports 31.82% AA Vals Index Model card Model card quotes: We use the Claude Code harness for Terminal-Bench 4.0 I am guessing AA & Vals index uses mini-swe-agent thus causing a low
中文全文翻译正在补齐，可先查看下方英文原文。
English / 原文
DeepSeek-V4.1-Flash-1017's Terminal bench 4.0 score: - Model card reported 42.3% - AA reports 34% - Vals index reports 31.82% AA Vals Index Model card Model card quotes: We use the Claude Code harness for Terminal-Bench 4.0 I am guessing AA & Vals index uses mini-swe-agent thus causing a low