DeepSeek-V4 preview version officially launched, providing a brand new experience of 1M ultra-long contextual memory.

On April 24th, DeepSeek announced that the preview version of the new series model DeepSeek-V4 is officially online and open source. DeepSeek-V4 has a million-word super long context and achieves leading capabilities in Agent ability, world knowledge, and reasoning performance in both domestic and open source fields. The model is divided into two versions based on size: deepseek-v4-flash and deepseek-v4-pro. Starting today, by logging into the official website or official App, you can interact with the latest DeepSeek-V4 and explore a new experience of 1M super long context memory. The API service has been updated synchronously, and you can call it by modifying the model_name to deepseek-v4-pro or deepseek-v4-flash. Compared to its predecessor models, the Agent ability of DeepSeek-V4-Pro has been significantly enhanced. In Agentic Coding evaluation, V4-Pro has reached the best level among current open source models and has also performed excellently in other Agent-related evaluations. Currently, DeepSeek-V4 has become the Agentic Coding model used by internal employees of the company, and feedback from evaluations shows that the user experience is better than Sonnet 4.5, and the delivery quality is close to Opus 4.6 non-thinking mode, but there is still some gap compared to Opus 4.6 thinking mode. It is reported that DeepSeek-V4 has pioneered a new attention mechanism, compressing at the token dimension and combining with DSA sparse attention (DeepSeek Sparse Attention), achieving globally leading long context capabilities while significantly reducing the requirements for computing and memory compared to traditional methods. From now on, 1M (one million) context will be a standard feature of all official DeepSeek services. V4-Pro and V4-Flash both support a maximum context length of 1M and support both non-thinking mode and thinking mode. Among them, the thinking mode supports the reasoning_effort parameter for setting the thinking intensity (high/max). For complex Agent scenarios, it is recommended to use the thinking mode and set the intensity to max. Currently, the DeepSeek API has synchronously launched V4-Pro and V4-Flash, supporting OpenAI ChatCompletions interface and Anthropic interface. When accessing the new model, the base_url remains the same, and the model parameter needs to be changed to deepseek-v4-pro or deepseek-v4-flash.