DeepSeek-V4.1-Flash · DeepSWE(v1.1; Resolved metric as printed in V4.1-Flash card) 66.2% · 附件证据(deepseek-v41-flash-ev-005)

DeepSeek-V4.1-Flash 的 DeepSWE · v1.1; Resolved metric as printed in V4.1-Flash card 官方出处。Pi v0.84.2 · 66.2%。见《DeepSeek-V4.1-Flash Technical Report》PDF p.35; Table 4; full header and target row。证据 deepseek-v41-flash-ev-005。

原文与定位

DeepSeek-V4.1-Flash · DeepSWE(v1.1; Resolved metric as printed in V4.1-Flash card),Pi v0.84.2配置,Resolved 66.2%(具体成绩公布时间未确认)。见《DeepSeek-V4.1-Flash Technical Report》PDF p.35; Table 4; full header and target row。 查看成绩条件

DeepSeek Harness Benchmark (Metric) | Claude Code | Codex | OpenCode | Pi | mini-SWE | Minimal | Standard | PTC | DeepSWE v1.1 (Resolved) | 69.8 | 65.6 | 65.5 | 66.2 | 74.2 | 72.6 | 70.5 | 67.6 |
官方材料
DeepSeek-V4.1-Flash Technical Report · PDF 第 35 页
定位
附件 · PDF p.35; Table 4; full header and target row
材料版本
Official V4.1-Flash report: 51/51 PDF pages read; Table 1, Table 3, Table 4, Table 5, multi-agent results and exact-label results were checked. CyberGym, SEC-Bench Pro and ExploitGym were retained as local audit only and excluded from regular mappings.
版本公布
未确认
披露时间
未确认 · 首发/后补未确认
采集时间
历史依据
暂无可靠历史版本依据
引用文本

DeepSeek-V4.1-Flash · DeepSWE(v1.1; Resolved metric as printed in V4.1-Flash card),Pi v0.84.2配置,Resolved 66.2%(具体成绩公布时间未确认)。官方出处 DeepSeek-V4.1-Flash Technical Report(附件,Official V4.1-Flash report: 51/51 PDF pages read; Table 1, Table 3, Table 4, Table 5, multi-agent results and exact-label results were checked. CyberGym, SEC-Bench Pro and ExploitGym were retained as local audit only and excluded from regular mappings.)PDF p.35; Table 4; full header and target row,https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash/resolve/main/DeepSeek_V41_Tech_Report.pdf?download=true#page=35。Benchlog https://benchlog.lekshome.cc/evidence/v41-flash/deepseek-v41-flash-ev-005/

关联成绩与条件

DeepSeek-V4.1-Flash · DeepSWE 成绩与配置
指标与成绩配置与评测条件证据关系
Resolved66.2%原文 66.2 · 单位 %
公布 未确认
Pi v0.84.2
设置
Pi v0.84.2 RPC mode with file and shell tools plus search extension. Max reasoning effort; Linux containers, temperature=1.0, top-p=0.95, 1M context and max_steps=500 as stated in the report note. This is an ordinary single-agent public scaffold run, not the Section 5.3.5 multi-agent experiment.
未知项
Public reproducible ordinary single-agent scaffold at explicit Max reasoning effort. Presentation uses the existing max mode; scaffold and version remain conditions and do not create a separate effort tier.
工具
允许工具
推理预算
Max
框架
未记录
协议
DeepSWE v1.1; Pi v0.84.2; N=8; Max reasoning_effort=100; temperature=1.0, top_p=0.95; 1M context; max_steps=500.
条件
Pi v0.84.2 scaffold; Max reasoning effort. 报告正文明确该指标为百分比。 Ordinary single-agent public scaffold; not a cross-version average or multi-agent run.

直接支持该成绩

在排名中查看此成绩