Skip to content
Seedance 2.0 Face is here — generate video from real-person reference photos.Try it now →
Back to Models
deepseekdeepseek·Apr 24, 2026

DeepSeek: DeepSeek V4 Pro

deepseek/deepseek-v4-pro
context
1.0M
Max Output
384K
Input / 1M
$1.74
Output / 1M
$3.48

About

DeepSeek V4 Pro is a large-scale Mixture-of-Experts model from DeepSeek with 1.6T total parameters and 49B activated parameters, supporting a 1M-token context window. Each MoE layer uses 384 routed experts plus 1 shared expert, with 6 experts active per token. It combines a hybrid attention mechanism (Compressed Sparse Attention and Heavily Compressed Attention) for efficient long-context processing and supports multiple reasoning modes to balance speed and depth. Reported benchmarks include 80.6% on SWE-bench Verified, 67.9% on Terminal-Bench 2.0, and 93.5% on LiveCodeBench. It is suited to full-codebase analysis, multi-step automation, and large-scale information synthesis.

Capabilities

Context Length
1.0M
Max Output
384K
Reasoning
Yes
In
text
Out
text

Benchmarks

View leaderboard
44.3#16 of 133
Intelligence Index
47.5#15 of 118
Coding Index

Reasoning & Knowledge

GPQA Diamond88.8%
HLE35.9%

Coding & Agentic

SciCode50.0%
Terminal-Bench Hard46.2%

Source: Artificial Analysis

Pricing

Full pricing
TypePrice / 1M tokens
Input$1.74
Output$3.48
Cache Read$0.145

API

API reference

OpenAI-compatible · Model ID deepseek/deepseek-v4-pro

curl https://api.elliotgate.com/v1/chat/completions \
  -H "Authorization: Bearer sk-omg-your-api-key" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek/deepseek-v4-pro",
    "messages": [{"role": "user", "content": "Hello!"}]
  }'

OFTEN COMPARED

DeepSeek: DeepSeek V4 Pro comparisons

Decide which model wins on the dimensions that matter for your workload — context, benchmarks, pricing, or serving latency.