Skip to main content

Your Privacy Choices

Copyright © 2026 NVIDIA Corporation

NVIDIA NIM | Try NVIDIA NIM APIs

deepseek-ai/deepseek-v4-pro

Deploy

Ready to scale? Choose your deployment path.

Partner EndpointsSelf-Hosted Deployments

Available Integrations

Deploy this model now on your endpoint provider of choice

Specifications

DeepSeek V4 scales to 1M-token context windows with efficient MoE architecture for coding tasks.

Moe
agentic
coding
reasoning

Provider: DeepSeek AI
Last Modified: 4 months ago
Context Length: 1M
Parameters: 2T
Input Modalities: Text
Output Modalities: Text

Model Availability

Free Endpoint: Deprecated
Partner Endpoint: Available
Download Available: Available
API calls in last 30 days: 7M