DeepSeek-V4 has achieved leading performance in Agent capabilities, world knowledge, and reasoning in both the domestic and open-source fields.

DeepSeek-V4 has achieved leading performance in Agent capabilities, world knowledge, and reasoning in both the domestic and open-source fields.

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

DeepSeek-V4 features million-character ultra-long context, leading domestic and open-source fields in Agent capabilities, world knowledge, and reasoning performance. The models are divided into two versions by size:

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

DeepSeek-V4-Pro: Performance on par with top closed-source models

  • Significantly improved Agent capability: Compared to previous models, DeepSeek-V4-Pro’s Agent capability has increased markedly. In Agentic Coding evaluation, V4-Pro has reached the best level among current open-source models and also excels in other agent-related evaluations. Currently, DeepSeek-V4 has become the Agentic Coding model used by company employees, and according to feedback, user experience is better than Sonnet 4.5, delivery quality is close to Opus 4.6 non-thinking mode, but there is still some gap compared to Opus 4.6 thinking mode.
  • Rich world knowledge: DeepSeek-V4-Pro is far ahead of other open-source models in world knowledge evaluations, only slightly behind top closed-source model Gemini-Pro-3.1.
  • World-class reasoning performance: In mathematics, STEM, and competitive coding evaluations, DeepSeek-V4-Pro surpasses all open-source models evaluated so far, achieving outstanding results comparable to world-class closed-source models.

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

Compared to DeepSeek-V4-Pro, DeepSeek-V4-Flash is slightly inferior in world knowledge reserves, but demonstrates similar reasoning ability. With smaller model parameters and activations, V4-Flash can provide faster and more economical API service.

In Agent evaluations, DeepSeek-V4-Flash matches DeepSeek-V4-Pro in simple tasks but still lags behind in more challenging tasks.

Structural innovation and ultra-high context efficiency
DeepSeek-V4 has pioneered a brand new attention mechanism, compressing along the token dimension, combined with DSA sparse attention (DeepSeek Sparse Attention), achieving world-leading long-context capability and greatly reducing computational and memory requirements compared to traditional methods. From now on, 1M (one million) context will be the standard for all official DeepSeek services.

Special Agent Capability Optimization

DeepSeek-V4 has been adapted and optimized for mainstream Agent products such as Claude Code, OpenClaw, OpenCode, and CodeBuddy, showing improvement in coding tasks, document generation tasks, and more. Below is an example of a PPT inner page generated by V4-Pro under a certain Agent framework:

API Access

Currently, DeepSeek API has launched V4-Pro and V4-Flash, supporting OpenAI ChatCompletions interface and Anthropic interface. When accessing the new models, the base_url remains unchanged; the model parameter should 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 mode and thinking mode. In thinking mode, reasoning_effort parameter can be set to adjust reasoning intensity (high/max). For complex Agent scenarios, it is recommended to use thinking mode and set intensity to max.

Final Words

"Not tempted by praise, nor fearful of slander, acting in accord with the way, upright and self-disciplined."

Thank you to every user for your trust and support; your recognition, suggestions, and expectations are the driving force for our continuous exploration and progress, enabling us to always hold true to our original intentions and focus on relentless innovation.

We will continue to uphold the principle of long-termism, moving forward steadily through attempts and reflections, striving to get closer to realizing AGI.

Risk Warning and DisclaimerThe market has risks, and investment needs caution. This article does not constitute personal investment advice, nor does it take into account the special investment goals, financial situations, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their specific circumstances. Investing based on this information is at your own risk.