DeepSeek-V4-Pro (Preview) · SWE-bench Verified(Version/task snapshot unspecified in V4 preview table; Resolved) 73.6% · 主文证据(ev-v4-pro-preview-018)

DeepSeek-V4-Pro (Preview) 的 SWE-bench Verified · Version/task snapshot unspecified in V4 preview table; Resolved 官方出处。DeepSeek-V4-Pro (Preview) Non-Think · 73.6%。见《DeepSeek-V4-Pro official Hugging Face model card》Official model card / table headed “| Benchmark (Metric) | V4-Flash Non-Think | V4-Flash High | V4-Flash Max | V4-Pro Non-Think | V4-Pro High | V4-Pro Max |” / complete row containing “| SWE Verified (Resolved) | 73.7 | 78.6 | 79.0 | 73.6 | 79.4 | **80.6** |”。证据 ev-v4-pro-preview-018。

原文与定位

DeepSeek-V4-Pro (Preview) · SWE-bench Verified(Version/task snapshot unspecified in V4 preview table; Resolved),DeepSeek-V4-Pro (Preview) Non-Think配置,Resolved 73.6%(具体成绩公布时间未确认)。见《DeepSeek-V4-Pro official Hugging Face model card》Official model card / table headed “| Benchmark (Metric) | V4-Flash Non-Think | V4-Flash High | V4-Flash Max | V4-Pro Non-Think | V4-Pro High | V4-Pro Max |” / complete row containing “| SWE Verified (Resolved) | 73.7 | 78.6 | 79.0 | 73.6 | 79.4 | **80.6** |”。 查看成绩条件

| Benchmark (Metric) | V4-Flash Non-Think | V4-Flash High | V4-Flash Max | V4-Pro Non-Think | V4-Pro High | V4-Pro Max | | SWE Verified (Resolved) | 73.7 | 78.6 | 79.0 | 73.6 | 79.4 | **80.6** |
官方材料
DeepSeek-V4-Pro official Hugging Face model card
定位
主文 · Official model card / table headed “| Benchmark (Metric) | V4-Flash Non-Think | V4-Flash High | V4-Flash Max | V4-Pro Non-Think | V4-Pro High | V4-Pro Max |” / complete row containing “| SWE Verified (Resolved) | 73.7 | 78.6 | 79.0 | 73.6 | 79.4 | **80.6** |”
材料版本
HF revision b5968e9190ef611bbf34a7229255be88a0e937c1; current card snapshot
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-Pro (Preview) · SWE-bench Verified(Version/task snapshot unspecified in V4 preview table; Resolved),DeepSeek-V4-Pro (Preview) Non-Think配置,Resolved 73.6%(具体成绩公布时间未确认)。官方出处 DeepSeek-V4-Pro official Hugging Face model card(主文,HF revision b5968e9190ef611bbf34a7229255be88a0e937c1; current card snapshot)Official model card / table headed “| Benchmark (Metric) | V4-Flash Non-Think | V4-Flash High | V4-Flash Max | V4-Pro Non-Think | V4-Pro High | V4-Pro Max |” / complete row containing “| SWE Verified (Resolved) | 73.7 | 78.6 | 79.0 | 73.6 | 79.4 | **80.6** |”,https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro。Benchlog https://benchlog.lekshome.cc/evidence/v4-pro-preview/ev-v4-pro-preview-018/

关联成绩与条件

DeepSeek-V4-Pro (Preview) · SWE-bench Verified 成绩与配置
指标与成绩配置与评测条件证据关系
Resolved73.6%原文 73.6 · 单位 %
公布 未确认
DeepSeek-V4-Pro (Preview) Non-Think
设置
Official mode label: Non-Think; the row-specific inference budget and sampling parameters are not stated in this model-card table.
未知项
Do not assume local recommended parameters are the evaluation settings.
工具
工具条件未记录
推理预算
未记录
框架
未记录
协议
V4 preview in-card comparison table; test-specific harness details are not fully stated.
条件
Official V4 preview mode comparison; Non-Think mode.

直接支持该成绩

在排名中查看此成绩