DeepSeek V4

DeepSeek-V4 Preview: Entering the Era of Million-Token Context for All

Today, the preview version of our new model series DeepSeek-V4 is officially launched and open-sourced.

DeepSeek-V4 features a million-token ultra-long context, achieving leading performance in Agent capabilities, world knowledge, and reasoning among domestic and open-source models. The model is available in two sizes:

Starting today, you can log in to the official websitechat.deepseek.comor the official App to chat with the latest DeepSeek-V4 and explore the new experience of 1M ultra-long context memory. The API service has been updated simultaneously; you can call it by changing the model_name to deepseek-v4-pro or deepseek-v4-flash.

DeepSeek-V4-Pro: Performance Comparable to Top Closed-Source Models

• Significantly improved Agent capabilities: Compared to the previous generation, DeepSeek-V4-Pro's Agent capabilities have been notably enhanced. In Agentic Coding evaluations, V4-Pro has reached the best level among current open-source models and performs excellently in other Agent-related evaluations. Currently, DeepSeek-V4 has become the Agentic Coding model used internally by employees, with feedback indicating a better user experience than Sonnet 4.5, delivery quality close to Opus 4.6 non-thinking mode, but still a gap compared to Opus 4.6 thinking mode.

• Rich world knowledge: In world knowledge evaluations, DeepSeek-V4-Pro significantly leads other open-source models, only slightly behind the top closed-source model Gemini-Pro-3.1.

• World-class reasoning performance: In evaluations of mathematics, STEM, and competitive coding, DeepSeek-V4-Pro surpasses all publicly evaluated open-source models, achieving excellent results comparable to top closed-source models.

DeepSeek-V4-Flash: A Faster and More Economical Choice

• Compared to DeepSeek-V4-Pro, DeepSeek-V4-Flash has slightly less world knowledge but demonstrates similar reasoning capabilities. Due to smaller model parameters and activation, V4-Flash offers faster and more economical API services.

• In Agent evaluations, DeepSeek-V4-Flash performs on par with DeepSeek-V4-Pro on simple tasks, but still lags on high-difficulty tasks.

DeepSeek-V4 introduces a novel attention mechanism that compresses tokens, combined with DSA sparse attention (DeepSeek Sparse Attention), achieving world-leading long-context capabilities while significantly reducing computational and memory requirements compared to traditional methods. From now on, 1M (one million) context will be standard for all official DeepSeek services.

DeepSeek-V4 has been adapted and optimized for mainstream Agent products such as Claude Code, OpenClaw, OpenCode, and CodeBuddy, with improvements in code tasks, document generation, and more. The image below shows an example of a PPT slide generated by V4-Pro in an Agent framework:

Currently, the DeepSeek API has simultaneously launched V4-Pro and V4-Flash, supporting the OpenAI ChatCompletions interface and the Anthropic interface. When accessing the new models, the base_url remains unchanged, and the model parameter needs to be changed to deepseek-v4-pro or deepseek-v4-flash.

V4-Pro and V4-Flash have a maximum context length of 1M, both supporting non-thinking and thinking modes, where the thinking mode supports the reasoning_effort parameter to set thinking intensity (high/max). For complex Agent scenarios, it is recommended to use thinking mode with intensity set to max. For model invocation and parameter adjustment methods, please refer to theThinking Mode Guide - API Documentation.

Note: The two model names deepseek-chat and deepseek-reasoner in the old API interface will be discontinued after three months (2026-07-24). During this period, these two model names point to the non-thinking and thinking modes of deepseek-v4-flash, respectively.

• DeepSeek-V4 Open-Source Weights - ModelScope

• DeepSeek-V4 Technical Report - arXiv

We thank every user for their trust and support. Your affirmation, suggestions, and expectations are the driving force for our relentless exploration and continuous progress, and they keep us true to our original aspiration of focused innovation.

We will always adhere to the principle of long-termism, steadily moving forward through experimentation and reflection, striving to get closer to the goal of achieving AGI.