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.
LatestDots Studio3 preview
Dots StudioAny-to-Any

dots3-note preview brings audio and vision to one model

An early build of a multimodal, long-context agentic system arrives on Hugging Face with support for both images and sound.

Aug 9, 2026
UpdateOther
dots3-note (preview)

dots-studio has published an early preview of dots3-note, a multimodal model designed to handle images, audio, and text within a single long-context, agentic framework. The release is available now on Hugging Face.

The model is described as "multimodal-any," meaning it aims to accept a range of input types rather than specializing in a single one. Vision-language capabilities sit alongside audio understanding, positioning dots3-note as a general-purpose system for tasks that blend perception with reasoning.

What we know

  • Multimodal support spanning vision and audio, plus text
  • A long-context, agentic design intended for multi-step tasks
  • Distributed under a non-standard "other" license
  • Published as a preview, so specifications may still shift

Key technical details remain unpublished at this stage. The record lists no confirmed parameter count, context length, or benchmark results, and the license is marked simply as "other" — worth checking carefully before any commercial use.

Why it matters: Combining audio and vision in one openly available agentic model is still relatively uncommon, and preview releases like this offer early hands-on access for developers experimenting with cross-modal workflows. As with any preview, expect the details to firm up as the family matures.

Sources

  • dots-studio/dots3-note-prev

    Hugging Face

    Visit
OlderQwen releases 2.4T-parameter open MoE with 95B activeQwen · Alibaba · Text / LLM · 3 months agoNewerMeta's Muse Glimmer 30B Targets Local Agentic CodingMeta AI · Vision-Language · 3 months ago

Get the model

Hugging Face

Specs

Context window524K tokens
Size576.9 GB
PrecisionBF16
ArchitectureDots3NoteForCausalLM
LicenseOTHER
Downloads4.1K
Likes296

Can you run it?

Needs a multi-GPU server — about 174.2 GB at 4-bit.

  • BF16 (as published)

    8×H100 80 GB node

    578.3 GB
  • 8-bit

    4×H100 80 GB · 512 GB Mac Studio

    306.7 GB
  • 4-bit

    4×H100 80 GB · 512 GB Mac Studio

    174.2 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

Any-to-AnyVision-Language

The Weekly Weights

Every open release that mattered, one email a week.

0 comments

No comments yet. Be the first to weigh in.

More in Any-to-Any

All Any-to-Any →
Sage S2.1 Flash
Internlm/ReasoningMulti-GPU server

InternLM's Sage S2.1 Flash targets scientific reasoning

The 320B mixture-of-experts model is built to reason over scientific evidence and act inside interactive environments.

Nov 15, 2026
Haru-4B
Higurashi Ai/Any-to-AnyRuns on a laptop

Haru-4B Brings Full-Duplex Speech to Six Languages

A compact 4-billion-parameter model handles real-time, two-way voice conversation across six languages in a single system.

Nov 12, 2026
LongCat 2.5
Meituan/Text / LLMDatacenter

Meituan Releases LongCat 2.5, an MIT-Licensed MoE

The Chinese delivery giant expands its open-model push with a multimodal, agentic mixture-of-experts system.

Nov 4, 2026