DeepSeek V4 Pro 0813 (on OpenRouter)
https://openrouter.ai/deepseek/deepseek-v4-pro-0813">DeepSeek V4 Pro 0813 (on OpenRouter)
The latest DeepSeek Pro model is now available, via API only.
I had to link to OpenRouter because DeepSeek don't have any obvious announcement page for their new model.
I haven't been able to confirm if they plan to release the open weights, but given the weights are available for both April's https://huggingface.co/deepseek-ai/DeepSeek-V4-Pro">deepseek-ai/DeepSeek-V4-Pro and July's https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-0731">deepseek-ai/DeepSeek-V4-Flash-0731 it seems likely.
Interestingly I got https://tools.simonwillison.net/markdown-svg-renderer#url=https%3A%2F%2Fgist.github.com%2Fsimonw%2Fc1108a380593547c2def5863bca63160">very different looking pelicans for the three different reasoning levels of low, medium, and high.
I've not noticed this kind of difference from any other model:
Low:
https://static.simonwillison.net/static/2026/deepseek-pro-low.png" />
Medium:
https://static.simonwillison.net/static/2026/deepseek-pro-medium.png" />
High:
https://static.simonwillison.net/static/2026/deepseek-pro-high.png" />
In terms of benchmarks... as far as I can tell those were released to the Official DeepSeek WeChat Group, then copied and pasted into https://www.reddit.com/r/LocalLLaMA/comments/1vmi0fg/removed_by_moderator/">a post on Reddit which was deleted by the moderators for being "low-effort", then copied into https://news.ycombinator.com/item?id=49274600#49275180">this ASCII-art table on Hacker News.
Tags: https://simonwillison.net/tags/ai">ai, https://simonwillison.net/tags/generative-ai">generative-ai, https://simonwillison.net/tags/llms">llms, https://simonwillison.net/tags/pelican-riding-a-bicycle">pelican-riding-a-bicycle, https://simonwillison.net/tags/deepseek">deepseek, https://simonwillison.net/tags/llm-release">llm-release, https://simonwillison.net/tags/ai-in-china">ai-in-china