The Open Weights
LatestModelsLeaderboardsCompanies
Subscribe
The Open Weights

The daily record of open-source AI. New model releases, leaderboards, and what's coming next — written for people who ship.

Refreshed every 12 hours

Discover

  • Latest releases
  • New today
  • Trending models

Browse

  • All models
  • Companies
  • Categories
  • Leaderboards

About

  • About
  • Editorial policy
  • Hardware estimates
  • RSS feed
  • llms.txt
  • Newsletter

© 2026 The Open Weights. An independent publication.

PrivacyTermsSMSAggregated by Claude · curated by humans.
LatestDeepSeekV4-Pro-0813
DeepSeekText / LLM

DeepSeek Releases V4-Pro-0813 With Open Weights

The Chinese lab pushes a higher-capability checkpoint of its V4 line to Hugging Face under a permissive MIT license.

Aug 13, 2026
NotableMIT
DeepSeek-V4-Pro-0813

DeepSeek has posted a new checkpoint, DeepSeek-V4-Pro-0813, to Hugging Face. Billed as a higher-capability member of the V4 Pro line, the model targets both general text generation and reasoning tasks, and it ships with downloadable weights rather than being locked behind an API.

The release continues DeepSeek's practice of using a mixture-of-experts (MoE) architecture, in which only a subset of the network's parameters activate for any given token. That design has become the lab's signature approach to squeezing strong performance out of large models while keeping inference costs manageable — a strategy that helped earlier DeepSeek releases punch above their weight against far larger dense competitors.

Why it matters

Permissive licensing is the headline here. DeepSeek is distributing the model under the MIT license, one of the most liberal terms available, which allows commercial use, modification, and redistribution with minimal restrictions. For teams evaluating open alternatives to proprietary frontier systems, that combination of open weights and a clean license lowers the barrier to experimentation and deployment.

A few practical caveats accompany the launch:

  • Detailed specifications such as total and active parameter counts and context length were not published in the release record.
  • Independent benchmarks are not yet available, so capability claims should be treated as provisional until third parties test the weights.

As always with a fresh checkpoint, the real test will come from the community. Expect quantized builds, fine-tunes, and head-to-head comparisons to surface quickly on the model's Hugging Face page as developers put V4-Pro through its paces.

Sources

  • deepseek-ai/DeepSeek-V4-Pro-0813

    Hugging Face

    Visit
OlderSuperwhisper's s1-mini polishes raw speech-to-text outputSuperwhisper · Speech → Text · 3 months agoNewerTencent's UI-Mate-27B targets desktop automationTencent · Vision-Language · 3 months ago

Get the model

Hugging Face

Specs

Context window1.0M tokens
Size1.7 TB
PrecisionINT8
ArchitectureDeepseekV4ForCausalLM
LicenseMIT
Downloads43.2K
Likes921

Can you run it?

Datacenter only — about 992 GB even at 4-bit.

  • FP8 (as published)

    Multi-node cluster

    1.7 TB
  • 4-bit

    Multi-node cluster

    992 GB
Your machine
GGUF builds
  • Mixture-of-experts: every expert must be loaded, so memory follows total parameters, not the active slice.
  • KV cache sized for 8,192 tokens of context; longer prompts need proportionally more.

Based on Hugging Face weights metadata. Assumes an 8K context and ~1 GB runtime overhead; actual needs vary. How we estimate


Modalities

Text / LLMReasoning
2 versions — view changelog

The Weekly Weights

Every open release that mattered, one email a week.

0 comments

No comments yet. Be the first to weigh in.

More from DeepSeek

All DeepSeek releases →
DeepSeek-V4.1-Flash
DeepSeek/Text / LLMDatacenter

DeepSeek Releases V4.1-Flash, an Open Multimodal MoE

The new mixture-of-experts model handles text, vision, and reasoning, and ships under a permissive MIT license.

Oct 17, 2026
DeepSeek-V4.1-Flash (1017)
DeepSeek/Vision-LanguageDatacenter

DeepSeek V4.1-Flash Adds Vision and Reasoning

The new MIT-licensed checkpoint folds image understanding and reasoning into DeepSeek's fast, mixture-of-experts Flash line.

Oct 17, 2026
DeepSeek-V4-Flash-Vision-Exp
DeepSeek/Vision-LanguageMulti-GPU server

DeepSeek adds vision to its V4 Flash line

An experimental, MIT-licensed vision-language model brings image understanding to DeepSeek's fast V4 Flash architecture.

Aug 31, 2026

More in Text / LLM

All Text / LLM →
Aion-1.0-Instruct-2.3B
Microsoft/Text / LLMRuns anywhere

Microsoft's Aion 1.0 Instruct is a compact 2.3B chat model

A small, multilingual instruction-tuned model aimed at chat and summarization lands on Hugging Face.

Nov 18, 2026
SenseNova-Loop-9B
SenseTime/Vision-LanguageRuns on a laptop

SenseTime Releases Looped 9B Multimodal Model

The company's new SenseNova-Loop-9B pairs vision-language capabilities with a recurrent looped-transformer design.

Nov 17, 2026
Hunyuan Hy4-flash
Tencent/Vision-LanguageMulti-GPU server

Tencent releases Hunyuan Hy4-flash multimodal MoE

The lightweight, mixture-of-experts model adds vision and text instruction-following to Tencent's open Hunyuan lineup.

Nov 17, 2026