# DeepSeek V4 Preview Release

🚀 **DeepSeek-V4 Preview** is officially live & open-sourced! Welcome to the era of cost-effective 1M context length.

🔹 **DeepSeek-V4-Pro:** 1.6T total / 49B active params. Performance rivaling the world's top closed-source models.

🔹 **DeepSeek-V4-Flash:** 284B total / 13B active params. Your fast, efficient, and economical choice.

Try it now at chat.deepseek.com via Expert Mode / Instant Mode. API is updated & available today!

📄 Tech Report: <https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro/blob/main/DeepSeek_V4.pdf>

🤗 Open Weights: <https://huggingface.co/collections/deepseek-ai/deepseek-v4>

![](https://api-docs.deepseek.com/img/v4-spec-en.png)

***

### DeepSeek-V4-Pro

🔹 **Enhanced Agentic Capabilities:** Open-source SOTA in Agentic Coding benchmarks. /&#x20;
🔹 **Rich World Knowledge:** Leads all current open models, trailing only Gemini-3.1-Pro. /&#x20;
🔹 **World-Class Reasoning:** Beats all current open models in Math/STEM/Coding, rivaling top closed-source models.

![](https://api-docs.deepseek.com/img/v4-benchmark.png)

***

### DeepSeek-V4-Flash

🔹 Reasoning capabilities closely approach V4-Pro. /&#x20;
🔹 Performs on par with V4-Pro on simple Agent tasks. /&#x20;
🔹 Smaller parameter size, faster response times, and highly cost-effective API pricing.

![](https://api-docs.deepseek.com/img/v4-benchmark-2.png)

***

### Structural Innovation & Ultra-High Context Efficiency

🔹 **Novel Attention:** Token-wise compression + DSA (DeepSeek Sparse Attention). /&#x20;
🔹 **Peak Efficiency:** World-leading long context with drastically reduced compute & memory costs. /&#x20;
🔹 **1M Standard:** 1M context is now the default across all official DeepSeek services.

![](https://api-docs.deepseek.com/img/v4-efficiency.png)

***

### Dedicated Optimizations for Agent Capabilities

🔹 DeepSeek-V4 is seamlessly integrated with leading AI agents like Claude Code, OpenClaw & OpenCode. /&#x20;
🔹 Already driving our in-house agentic coding at DeepSeek. /&#x20;

The figure below showcases a sample PDF generated by DeepSeek-V4-Pro.

![](https://api-docs.deepseek.com/img/v4-ppt-en.png)

***

### API is Available Today!

🔹 Keep base\_url, just update model to deepseek-v4-pro or deepseek-v4-flash. /&#x20;
🔹 Supports OpenAI ChatCompletions & Anthropic APIs. /&#x20;
🔹 Both models support 1M context & dual modes (Thinking / Non-Thinking): [Thinking Mode guide](https://api-docs.deepseek.com/guides/thinking_mode.md)

⚠️ Note: deepseek-chat & deepseek-reasoner will be fully retired and inaccessible after Jul 24th, 2026, 15:59 (UTC Time). (Currently routing to deepseek-v4-flash non-thinking/thinking).

![](https://api-docs.deepseek.com/img/v4-price-en.png)

***

🔹 Amid recent attention, a quick reminder: please rely only on our official accounts for DeepSeek news. Statements from other channels do not reflect our views. /&#x20;
🔹 Thank you for your continued trust. We remain committed to longtermism, advancing steadily toward our ultimate goal of AGI.