DeepSeek V4 Pro
Flagshipby DeepSeek
DeepSeek V4 Pro is a 1.6 trillion parameter Mixture-of-Experts model with 49 billion active parameters per token, trained on 33 trillion tokens. Released under the MIT license with a native 1M context window, it delivers benchmark scores within striking distance of frontier proprietary models at a fraction of the cost.
Input Price
$1.74
per 1M tokens
Output Price
$3.48
per 1M tokens
Context Window
1M
tokens
Released
2026-04
Open source
Capabilities
Key Strengths
- ✓1.6T total parameters
- ✓MIT open source license
- ✓1M native context
- ✓Near-frontier benchmarks at budget pricing
Best For
- ▸Self-hosted deployments
- ▸Cost-sensitive production
- ▸Code generation
- ▸Long document processing
Benchmark Scores
| Benchmark | Score | Description |
|---|---|---|
| MMLU-Pro | 91.5 | General knowledge and reasoning across 57 subjects |
| HumanEval | 94.8 | Python code generation and problem solving |
| GPQA Diamond | 73.1 | Graduate-level science questions verified by domain experts |
| MATH | 92.4 | Competition-level mathematics problems |
| SWE-bench | 63.8 | Real-world software engineering tasks from GitHub issues |
Scores sourced from public benchmark datasets. See full benchmark leaderboard for all models.
Pricing Details
Input tokens
$1.74
per 1M tokens
Output tokens
$3.48
per 1M tokens
Estimated cost per 1K requests
$3.48
~1K input + ~500 output tokens avg
Prices are subject to change. Check the official documentation for current pricing. See the cost calculator for detailed estimates.
Open Source Model
DeepSeek V4 Pro is free to download and self-host under the MIT. Hosted API pricing varies by provider (e.g., Together, Fireworks, Groq). See our open source LLM guide for deployment options.