DeepSeek V4 Pro 0813
What happened
DeepSeek released a new text‑generation model, DeepSeek V4 Pro 0813, with 1.7 trillion parameters. The model is hosted on Hugging Face and can be downloaded or run via several inference providers, including Together, Novita, DeepInfra, and Baseten. It supports conversational tasks, offers tool‑calling capabilities, and provides structured output on some platforms. Performance metrics show a Terminalbench 2.1 score of 87.9, a DeepSWE score of 62.7, and a Toolathlon score of 74.1. The model is available under an MIT license and is compatible with 8‑bit and FP8 inference.
Reality check
An automated, sceptical read of the sourceIncremental. The model adds a large‑parameter variant with tool‑calling, but follows the existing DeepSeek V4 line and similar offerings in the field.
Is this a big deal?
Get stories like this every morning
The day's AI news in plain English, in your inbox at 8:30 IST. Free.
More from DeepSeek
DeepSeek V4 Pro 0813DeepSeek · 14 Nov 2026DeepSeek UltrafastDeepSeek · 13 Nov 2026DeepInfer: High-performance inference framework for DeepSeek modelsDeepSeek · 20 Oct 2026DeepSeek banned from U.S. federal civilian agencies under new DHS orderDeepSeek · 19 Oct 2026DeepSeek V4.1 Flash 1017 (Max)DeepSeek · 19 Oct 2026This page was collected from huggingface.co and summarised automatically by AIPulsee's pipeline. See how it runs.