DeepSeek V4 Pro 0813

DeepSeekMCP Tool CallReasoning
deepseek/deepseek-v4-pro-0813Open Weights

DeepSeek V4 Pro snapshot with million-token context and support for thinking and non-thinking modes

Benchmark Pricing$2.00 / 1M in$4.00 / 1M out$0.50 cache read
Context: 1,000,000 tokens•Max Output: 384,000 tokens•Released: 2026-08-13

TECHNICAL SPECIFICATIONS

Context Window1,000,0001M tokens
Max Output384,000384K tokens
LicenseMITSelf-hostable
Input Modalities
Text
Function / Tool CallingSupported (MCP Ready)
Structured OutputsGuaranteed Schema

MODEL CONTEXT PROTOCOL (MCP) READINESS

✓ DeepSeek V4 Pro 0813 natively supports OpenAI / Anthropic format tool calling and is fully compatible with MCP servers and agent loops.

MCP Context Window Recommendations

With a context capacity of 1,000,000 tokens, this model can handle extremely large MCP resource manifests, complete database schemas, and multiple active server tool definitions simultaneously without degradation.

STANDARDIZED BENCHMARKS & EVALS

Humanity's Last Examscore
42.7
Humanity's Last Examscore
60
Terminal-Benchscore
87.9
NL2Reposcore
61.5
CyberGymscore
83.3
DeepSWEscore
62.7
Toolathlon-Verifiedscore
74.1
Agents' Last Examscore
25.7
AutomationBenchscore
31.8
DSBench-FullStackscore
71.1
DSBench-Hardscore
67.2

API CALL INVOCATION EXAMPLE

// Example: Calling DeepSeek V4 Pro 0813 via OpenAI SDK
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.DEEPSEEK_API_KEY,
});

const response = await client.chat.completions.create({
  model: "deepseek/deepseek-v4-pro-0813",
  messages: [{ role: "user", content: "Hello! Can you inspect MCP tools?" }],
});

console.log(response.choices[0].message.content);

PRICING BY PROVIDER

15 providers
Cortecsdeepseek-v4-pro-0813
$2.00 / $4.00$0.50 cache
NanoGPTdeepseek/deepseek-v4-pro-0813
$0.66 / $1.98$0.02 cache
RunInfradeepseek-ai/DeepSeek-V4-Pro-0813
$0.60 / $1.90$0.03 cache
Charm Hyperdeepseek-v4-pro-0813
$1.44 / $4.31$0.05 cache
Novita AIdeepseek/deepseek-v4-pro-0813
$1.32 / $3.96$0.04 cache
Vivgriddeepseek-v4-pro-0813
$1.35 / $3.00$0.05 cache
IteraComputedeepseek/deepseek-v4-pro-0813
$0.79 / $2.38$0.03 cache
CoreWeavedeepseek-ai/DeepSeek-V4-Pro-0813
$1.31 / $3.96$0.04 cache
AI/ML APIdeepseek/deepseek-v4-pro-0813
$2.39 / $4.79$0.20 cache
Boundlessdeepseek-v4-pro-0813
$1.04 / $2.08$0.04 cache
Volcengine Arkdeepseek-v4-pro-ga-260813
$1.34 / $4.01$0.04 cache
Venice AIdeepseek-v4-pro-0813
$1.65 / $4.95$0.17 cache
AIHubMixdeepseek-v4-pro-0813
$0.69 / $2.08$0.02 cache
Basetendeepseek-ai/DeepSeek-V4-Pro-0813
$1.32 / $3.96$0.13 cache
Vercel AI Gatewaydeepseek/deepseek-v4-pro-0813
$0.66 / $1.98$0.07 cache

Rates in USD per 1M tokens. Rates synchronized from models.dev public catalog.

Prompt Caching Active

This model supports prompt caching at $0.50 / 1M tokens, saving up to 75% on repetitive MCP tool definitions and system instructions.

MODEL RESOURCES

View DeepSeek HubHugging Face Weights

More Models by DeepSeek

Alternative and family models from the same lab

View all DeepSeek models →
DeepSeek V4.1 Flash 1017
deepseek/deepseek-v4.1-flash-1017•DeepSeek
1M ctx

October 17 update of DeepSeek V4.1 Flash with improved reasoning, coding, agentic tasks, and visual understanding

384K out maxVisionHumanity's Last Exam: 42.6 (unit not reported)
In: $0.15 • Out: $0.60
DeepSeek
DeepSeek V4 Pro 0423
deepseek/deepseek-v4-pro-0423•DeepSeek
1M ctx

DeepSeek V4 Pro initial snapshot with million-token context and support for thinking and non-thinking modes

384K out max
In: $1.32 • Out: $3.96
DeepSeek
DeepSeek V4 Pro
deepseek/deepseek-v4-pro•DeepSeek
1M ctx

Open MoE flagship with million-token context for coding and long agent runs

384K out maxSWE-Bench Verified: 80.6 resolved
In: $0.66 • Out: $1.98
DeepSeek