models/deepseek/deepseek-v4-pro
D
DeepSeek·active

DeepSeek V4 Pro

DeepSeek's mid-tier model. Long-context specialist with 1.0M window.

Overall score
4.46
/5.00 · ranked #40
Input
$0.435
per 1M tokens
Output
$0.870
per 1M tokens
Context
1.0M
tokens
Blended
$0.761
3:1 out:in ratio

Price drops, new benchmarks, model updates. Stay current on DeepSeek V4 Pro.

One email per change. Unsubscribe anytime.

modelpicker.aipowered by live benchmark data

Scores by test

Methodology →
Structured Output
5.0
Strategic Analysis
5.0
Constrained Rewriting
4.0
Creative Problem Solving
5.0
Tool Calling
4.0
Faithfulness
5.0
Classification
4.0
Long Context
4.0
Safety Calibration
2.0
Persona Consistency
5.0
Agentic Planning
5.0
Multilingual
5.0
Tabular Data
5.0
SWE-bench Verified
77.6
AIME 2025
96.7
GPQA Diamond
89.6
SciCode
50.0
Epoch Capabilities Index (ECI)
148.9

What you need to know

DeepSeek V4 Pro is a high-performance open-weight model optimized for complex logic and structured data. It excels in agentic planning, strategic analysis, and structured output, achieving perfect 5/5 internal scores across these domains. Its ability to handle tabular data and maintain persona consistency makes it a strong candidate for autonomous workflows and data-heavy applications.

The model provides significant value relative to its performance rank. With a blended cost of $0.761/MTok, it offers top-10 overall performance at a price point that is highly competitive for its capability tier. The 1.0M context window allows for massive data ingestion, though its internal long-context score of 4/5 suggests a slight drop in reliability compared to its perfect scores in logic and structure.

A critical trade-off is the model's safety calibration, which is its lowest metric at 2/5. This indicates a higher likelihood of generating unfiltered or non-compliant responses, requiring developers to implement robust external guardrails. While it performs well in most linguistic tasks, it is slightly less effective at constrained rewriting and classification than it is at creative problem solving.

Use this model if you need an affordable, open-weight solution for agentic workflows, complex strategic planning, or high-precision structured data generation. Skip this model if your application requires strict built-in safety filters or highly constrained text transformations.

Strengths — Top 3

Structured Output5.0/5.0
Strategic Analysis5.0/5.0
Creative Problem Solving5.0/5.0

Relative weaknesses — Bottom 3

Safety Calibration2.0/5.0
Constrained Rewriting4.0/5.0
Tool Calling4.0/5.0

Similar models

NNVIDIA: Nemotron 3 Ultra$2.854.46QQwen 3.7 Max$3.694.62GGemini 3.5 Flash$7.134.46QQwen: Qwen3.7 Flash$0.1054.54