DeepSeek V4 Pro 0813
开始使用
API 接入
Merge Gateway API文档
目录部署标识：deepseek/deepseek-v4-pro-0813。具体区域、权限与价格以服务商为准。
Kilo Gateway API文档
目录部署标识：deepseek/deepseek-v4-pro-0813。具体区域、权限与价格以服务商为准。
Novita AI API文档
目录部署标识：deepseek/deepseek-v4-pro-0813。具体区域、权限与价格以服务商为准。
TensorX API文档
目录部署标识：deepseek/deepseek-v4-pro-0813。具体区域、权限与价格以服务商为准。
IteraCompute API文档
目录部署标识：deepseek/deepseek-v4-pro-0813。具体区域、权限与价格以服务商为准。
OrcaRouter API文档
目录部署标识：deepseek/deepseek-v4-pro-0813。具体区域、权限与价格以服务商为准。
NanoGPT API文档
目录部署标识：deepseek/deepseek-v4-pro-0813。具体区域、权限与价格以服务商为准。
Vercel AI Gateway API文档
目录部署标识：deepseek/deepseek-v4-pro-0813。具体区域、权限与价格以服务商为准。
AI/ML API API文档
目录部署标识：deepseek/deepseek-v4-pro-0813。具体区域、权限与价格以服务商为准。
Ofox API文档
目录部署标识：deepseek/deepseek-v4-pro-0813。具体区域、权限与价格以服务商为准。
Standard Compute API文档
目录部署标识：deepseek/deepseek-v4-pro-0813。具体区域、权限与价格以服务商为准。
Arcee API文档
目录部署标识：deepseek/deepseek-v4-pro-0813。具体区域、权限与价格以服务商为准。
OpenRouter API文档
目录部署标识：deepseek/deepseek-v4-pro-0813。具体区域、权限与价格以服务商为准。
公开成绩与计算依据
综合参考分根据已公开的评测资料整理。查看发布日期依据
综合参考分 50.9 · 17项基准
推理求解 56.3
GPQA Diamond · 0.928282828282828 比例
Artificial Analysis · 独立评测 · Diamond
DeepSeek V4 Pro 0813 (Max) · Current AA standalone protocol. Status verified in aa-method-current.md section gpqa-diamond
Terminal-Bench-Science · 4.29%
Vals AI · 独立评测 · 0.1.0
DeepSeek V4 Pro 0813 / Mini-SWE-agent · Current displayed evaluator row. Mini-SWE-agent. Scores are not pooled across harnesses; source methodology and repeat counts retained.
HLE · 0.195 比例
Artificial Analysis · 独立评测 · Diamond
DeepSeek V4 Pro 0813 (Max) · Artificial Analysis current published protocol; hle-diamond
CritPt · 0.235 比例
Artificial Analysis · 独立评测 · 2 Hard
DeepSeek V4 Pro 0813 (Max) · Artificial Analysis current published protocol; critpt-v2-hard
工具与任务执行 57.5
APEX-Agents · 32.4 %
huggingface.co · 官方发布比较 · current published task suite
DeepSeek V4 Pro 0813 / official Hy4 comparison, APEX-Agents pass@1 · Original numeric labels visually read in official image; APEX-Agents; Official Hy4 preview blue series; not Hy3 cyan sub-series. JPEG read at decoder 1/8 scale, original SHA-256 preserved.
tau3-Banking · 0.395876288659794 比例
Artificial Analysis · 独立评测 · 1.0.1
DeepSeek V4 Pro 0813 (Max) · Current AA standalone protocol. Status verified in aa-method-current.md section tau3-banking
GDPval-AA · 1457.37 分
Artificial Analysis · 独立评测 · 2.1
DeepSeek V4 Pro 0813 (Max) · Artificial Analysis current published protocol; gdpval-aa-normalized
Terminal-Bench · 14.14%
Vals AI · 独立评测 · 4.0
DeepSeek V4 Pro 0813 / Mini-SWE-agent · Current displayed evaluator row. Mini-SWE-agent. Scores are not pooled across harnesses; source methodology and repeat counts retained.
AutomationBench · 0.5671165546344002 比例
Artificial Analysis · 独立评测 · current published task suite
DeepSeek V4 Pro 0813 (Max) · Artificial Analysis current published protocol; automationbench-aa
AA-Briefcase · 1253.16 分
Artificial Analysis · 独立评测 · 1.1
DeepSeek V4 Pro 0813 (Max) · Artificial Analysis current published protocol; aa-briefcase
EnterpriseOps-Gym · 0.4959713518352731 比例
Artificial Analysis · 独立评测 · current published task suite
DeepSeek V4 Pro 0813 (Max) · Artificial Analysis current published protocol; enterprise-ops-gym-aa
编程 38.5
FrontierCode · 0.170742929292929 比例
Artificial Analysis · 独立评测 · 2.0
DeepSeek V4 Pro 0813 (Max) · Artificial Analysis current published protocol; frontiercode-2-0
Vals Code Migration · 51.99%
Vals AI · 独立评测 · 1.1
DeepSeek V4 Pro 0813 / Published Code Migration 1.1 sandbox · Current displayed evaluator row. Published Code Migration 1.1 sandbox. Scores are not pooled across harnesses; source methodology and repeat counts retained.
DeepSWE · 62.7
huggingface.co · 官方发布比较 · 1.1
DS-V4-Pro · Published page deepseek-v41-flash; row DeepSWE v1.1 (Resolved); preserve page footnotes and task-specific settings
语言与信息处理 48.4
GDP.pdf · 0.114 比例
Artificial Analysis · 独立评测 · current published task suite
DeepSeek V4 Pro 0813 (Max) · Artificial Analysis current published protocol; gdp-pdf
SimpleQA Verified · 0.5290581162324649 比例
Epoch AI · 独立评测 · run 1.2.0
deepseek-v4-pro-0813_max · epoch-simpleqa-reviewed-method-20261029-task-version-1.2.0 · 2026-08-27公开
AA-Omniscience · 0.452 比例
Artificial Analysis · 独立评测 · 1.1
DeepSeek V4 Pro 0813 (Max) · Artificial Analysis current published protocol; omniscience-accuracy
AA-Omniscience · 0.05900243309002429 比例
Artificial Analysis · 独立评测 · 1.1
DeepSeek V4 Pro 0813 (Max) · Artificial Analysis current published protocol; omniscience-non-hallucination
资料来源
- 版本日期来源 · 版本资料 · 2026-10-29查阅
- Vals AI · independent_evaluation · 2026-10-29查阅
- huggingface.co · official_release_comparison · 2026-10-29查阅
- Vals AI · independent_evaluation · 2026-10-29查阅
- Vals AI · independent_evaluation · 2026-10-29查阅
- huggingface.co · official_release_comparison · 2026-10-29查阅
- Epoch AI · independent_evaluation · 2026-10-29查阅