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.
LatestMeituan1.6
MeituanImage → Video

Meituan's LongCat Adds Audio-Driven Avatar Video

The LongCat team releases an image-to-video model that animates still portraits from audio input.

Nov 13, 2026
NotableOther
LongCat-Video-Avatar-1.6

Meituan's LongCat team has published LongCat-Video-Avatar-1.6, an audio-driven avatar video generation model now available on Hugging Face. The model takes a still image and produces animated video, placing it in the fast-moving category of talking-head and digital-avatar generation.

The release is listed with an image-to-video primary modality, meaning it can take a source portrait and synthesize motion — the kind of lip-sync and expression animation that pairs a static image with an audio track. That makes it a candidate for use cases like virtual presenters, dubbing, and conversational interfaces.

Why it matters

Avatar video is one of the most commercially visible corners of generative media, and much of the leading work has stayed behind proprietary APIs. A major Chinese consumer-tech company putting a LongCat avatar model into open distribution gives developers something concrete to test and build on, rather than a closed service.

A few practical notes for anyone evaluating it:

  • The weights ship under a custom "other" license, so terms should be reviewed before commercial use.
  • Parameter count, context details, and benchmark figures were not disclosed in the release record.
  • The 1.6 version label suggests ongoing iteration within the broader LongCat-Video effort.

As always with avatar generation, the capability raises familiar questions about consent and synthetic likeness. For now, the model is available to pull and experiment with directly from the LongCat repository on Hugging Face.

Sources

  • meituan-longcat/LongCat-Video-Avatar-1.6

    Hugging Face

    Visit
OlderHaru-4B Brings Full-Duplex Speech to Six LanguagesHigurashi Ai · Any-to-Any · last weekNewerInternLM's Sage S2.1 Flash targets scientific reasoningInternlm · Reasoning · 5 days ago

Get the model

Hugging Face

Specs

LicenseOTHER
Downloads3.9K
Likes911

Modalities

Image → Video
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 Meituan

All Meituan releases →
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
LongCat-Flash-Lite-Sparse
Meituan/Text / LLMWorkstation GPU

Meituan Ships a Lighter, Sparser LongCat-Flash

The food-delivery giant's newest open model trims its mixture-of-experts design for more efficient inference under an MIT license.

Jul 31, 2026
LongCat 2.5
Meituan/Text / LLMDatacenter

Meituan releases LongCat-2.0 language model

The Chinese delivery giant continues its push into open AI with a new text model on Hugging Face.

Jun 30, 2026

More in Image → Video

All Image → Video →
NVIDIA/Image → Video

NVIDIA Releases Cosmos 3 for World Simulation

The new open models push NVIDIA's world-foundation effort toward interactive, real-time video generation.

Nov 5, 2026
NVIDIA/Text → Video

NVIDIA's Cosmos 3 Brings Open Interactive World Models

The new open release pushes beyond passive video generation toward simulated worlds that respond to input.

Nov 5, 2026
Hy-Nomas
Tencent/Text → Video

Tencent Releases Hy-Nomas, an Interactive Video World Model

The open-weights model pairs text- and image-to-video generation with action-aware causal attention and explicit camera control.

Oct 21, 2026