LongCat Video Avatar 1.6: what it is, how to run it, what to check first
LongCat Video Avatar 1.6 is Meituan's open-source model that turns speech audio, with an optional photo, into talking-person video, which the makers say can run past 10 minutes; this site currently delivers images only, not video, but its photo editor, a different model, edits one photo in about 37 seconds for 10 credits. Free with sign-up: 10 credits = 1 image, then from $0.13 per image.

Your photoStyle changes (cartoon, 3D, anime) are not reliable yet. If the result looks unchanged, try again or describe the change more specifically.
One-tap edits
- Not LongCat, not affiliated with Meituan
- Images only: this site does not make video
- Free with sign-up: 10 credits = 1 image
- Hugging Face repo created (HF API); Meituan’s README news line is dated Nov 15
- Nov 13, 2026
- licence of the model weights (model card)
- MIT
- BF16 base-model files; the INT8 set is 15.9 GB (our sums)
- 31.7 GB
- clip length the makers claim; not tested
- > 10 min
What LongCat Video Avatar 1.6 is
From the Hugging Face model card
- The Meituan LongCat Team calls it the latest upgrade to its open-source framework for audio-driven human video generation, building on LongCat-Video-Avatar 1.5. The card lists the base model as meituan-longcat/LongCat-Video.Source: Hugging Face model card ·
- Inputs, from the card’s tags and demo commands: speech audio with text (audio-text-to-video), speech audio with a reference image (audio-image-text-to-video), and audio-driven continuation of a clip. The demos cover single-person and multi-person animation.Source: Hugging Face model card ·
- Licence: the model weights are MIT. The card adds that this grants no rights to Meituan trademarks or patents, and says the model has not been comprehensively evaluated for every downstream use.Source: Hugging Face model card ·
- Languages: the card tags English and Chinese and notes that performance varies across languages, with no per-language results.Source: Hugging Face model card ·
- Dates: the model card gives no release date. The Hugging Face repo was created on 2026-11-13, and the README of the LongCat-Video repo has a news line dated November 15, 2026 for this release; ComfyUI Wiki dates it November 13.Source: Hugging Face API and GitHub README ·
The makers’ claims, not tested by us
- Version 1.6 is said to give more natural lip movement, steadier long videos (the card says videos exceeding 10 minutes while keeping identity and quality) and faster end-to-end inference. The card and the project page give no speed numbers.Source: Hugging Face model card ·
- A human-preference study run by the makers counts 33,051 valid votes over 15 pairings of six systems. It puts 1.6 second, with a Bradley–Terry Elo of 1068.1, behind OmniHuman-2.0 and ahead of 1.5, HeyGen, OmniHuman-1.5 and Kling Avatar 2.0. We could not see the method, and it is not an independent test.Source: Hugging Face model card ·
A licence question to settle before commercial use
- The weights are MIT, but the Ethical Considerations section at the bottom of the 1.6 project page says the generated content is for academic use only and that commercial use is not permitted. We could not tell whether that line is a content notice or a limit on use, and the two do not read the same. Ask Meituan (the card lists longcat-team@meituan.com) before you ship anything commercial. This is not legal advice.Source: LongCat-Video-Avatar 1.6 project page ·
How to run it, from the makers’ README
- Clone the LongCat-Video repo, make a Python 3.10 conda environment, install torch 2.6.0 (CUDA 12.4 build) and flash_attn 2.7.4.post1, install requirements.txt and requirements_avatar.txt, then download the weights with huggingface-cli. The demo scripts run with torchrun; the example commands use 2 GPUs with context parallelism.Source: GitHub README ·
- Options named in the README: FlashAttention 3 on Hopper GPUs (H20, H100, H200, H800), optional FP8 (needs Transformer Engine) or weight-only INT8, and the 480P and 720P settings of the avatar demos. The README states no GPU or memory minimum.Source: GitHub README ·
- ComfyUI: the model card does not mention it. ComfyUI Wiki, a community news site that says it is separate from official ComfyUI, reported that the community pack for LongCat avatars had not been updated for 1.6, so the official Python pipeline is the route it names.Source: ComfyUI Wiki (third party) ·
On this site
- LongCat Video, the base model that Avatar 1.6 is built on, has its own page here: what it is, where to run it, and fal.ai’s prices per second.Source: LongCat Video: what it is and where to run it
- To change a still photo instead, type the change you want in the AI photo editor: up to 400 characters, about 37 seconds, one image out.Source: AI photo editor: type the change
- Another video-model explainer on this site, read the same way: what the maker says, what we could not confirm.Source: Seedance 2.5: what it is
- All of this site’s photo effects, from restyles to holiday portraits, in one list.Source: All AI photo effects
Not confirmed
- Parameter count, output resolution and frame rate on the 1.6 card, maximum length of one segment, and any speed figure: the card states none of these. The README’s 480P and 720P line sits in tips shared with 1.5.Source: Hugging Face model card ·
- Hosted versions: the card says no inference provider runs it. fal.ai lists LongCat avatar endpoints in search results, but its pages returned 404 to us, so we quote no price and do not know which Avatar version they run.Source: Hugging Face model card ·
- The technical report (a PDF in the LongCat-Video repo) was not read by us, and the card’s arXiv fields are blank. Any figure that appears only there is not on this page.Source: GitHub README ·
Sizes
LongCat Video Avatar 1.6 download sizes, and this site
| - | Size on Hugging Face | What it is |
|---|---|---|
| base_model (6 files) | 31.71 GB | The full-precision (BF16) video model |
| base_model_int8 (4 files) | 15.88 GB | Optional weight-only INT8 version, switched on with --use_int8 |
| loraA and loraB | 2.52 GB each | Two adapters: DMD (loraA) and OPSD (loraB), mixed during sampling |
| This site | Nothing to download | One photo edited in about 37 seconds, 10 credits (1024 × 1024); no video |
Where to run it
How to run LongCat Video Avatar 1.6
Settle the licence question first
The model card says MIT for the weights, while the project page says generated content is for academic use only. For anything commercial, ask Meituan and keep the answer.
On your own GPUs: the open weights
Follow the GitHub README: Python 3.10, torch 2.6.0, flash_attn, the two requirements files, then huggingface-cli for the weights and a torchrun demo. The README’s examples use two GPUs, and it states no minimum, so test on a short clip first.
In ComfyUI or on a hosted API: not confirmed
We found no confirmed ComfyUI path for 1.6 and could not open any hosted listing with a price. Check the model card for news before you plan around either.
Need a still image instead? Use this site
This site currently delivers images only, not video. Upload one photo and describe the change in the editor, up to 400 characters.
Prompts
Photo edit prompts for this site’s editor
These prompts are for this site’s photo editor, which edits still images with a different model from LongCat Video Avatar 1.6. We did not run them for this page, and they do not make video or talking avatars.
One prompt per one-tap edit
Remove background clutter
Edit my uploaded photo: remove distracting clutter from the background (stray objects, litter, bins, cables, signs and people in the background) and fill those areas naturally with what is around them. Keep the framing as it is. Keep every face, expression, pose and outfit exactly the same. No text, logos or watermarks.
Plain studio background
Edit my uploaded photo: replace the background with a plain, softly lit light-grey studio backdrop with natural soft shadows. Keep the framing as it is. Keep every face, expression, pose and outfit exactly the same. No text, logos or watermarks.
Professional headshot
Edit my uploaded photo into a clean professional headshot: a smooth neutral grey-blue background with a gentle gradient, soft even studio light on the face, the same crop or a slightly tighter head-and-shoulders crop. Keep every face, expression, pose and outfit exactly the same. No text, logos or watermarks.
Fix the lighting
Edit my uploaded photo: fix the lighting only. Balance the exposure, lift dark shadows on faces, tame blown-out highlights and correct the white balance so skin tones look natural. Keep every face, expression, pose and outfit exactly the same. No text, logos or watermarks.
Warm golden-hour light
Edit my uploaded photo: change only the light to warm, low golden-hour sunlight from one side, with soft long shadows and a warm glow, as if taken shortly before sunset. Keep the scene and framing. Keep every face, expression, pose and outfit exactly the same. No text, logos or watermarks.
Classic black and white
Edit my uploaded photo into a classic black and white photograph with rich contrast, deep blacks, clean whites and fine film grain. Keep the composition exactly as it is. Keep every face, expression, pose and outfit exactly the same. No text, logos or watermarks.
Watercolor painting
Repaint my uploaded photo as a soft watercolor painting on textured white paper, with loose washes and visible brush edges, keeping the same composition. Keep every person and pet recognizable, and do not add or remove anyone. No text, logos or watermarks.
Sharpen and clean up
Edit my uploaded photo: gently sharpen it, reduce noise and compression artifacts, and recover fine detail in hair, fabric and edges, without a plastic or over-smoothed look. Keep every face, expression, pose and outfit exactly the same. No text, logos or watermarks.
Pricing on this site
Credit packs or a monthly plan.
Credits come in one-time packs ($4.99 for 50 credits, $9.99 for 120 credits, $19.99 for 300 credits); each image costs 10 credits. Or pick a plan from $9.90 a month (500 credits), cancel anytime.
Refund policy- Basic · $9.90/month500 credits a month · about 50 images · $0.20 per image
- Pro · $19.90/month1500 credits a month · about 150 images · $0.13 per image
Billed every month and renews automatically. Cancel anytime in Settings; you keep the plan until the end of the month you paid for. Each month's credits are for that month; unused ones expire.
Refund policyPhoto tools on this site
Edit, restore and restyle your own photo with this site’s image model
- Alternatives and AI toolsOpen
MiniMax H3 ImageOpen
LongCat VideoOpen
Adobe Firefly alternativeOpen
EasyFaceSwap alternativeOpen
Vidwud face swap alternativeOpen
MeiGen AI alternativeOpen
SnapGen AI alternativeOpen
Canva headshot generator alternativeOpen
QuickAIHeadshots alternativeOpen
Betterpic alternativeOpen
Everygen alternativeOpen
PhotoCraft alternativeOpen
Free clipping AIOpen
What is KixyOpen
OpenShorts alternativeOpen
Snazo alternativeOpen
Vikas edit alternativeOpen
Seedream 5.5Open
PromptSeen alternativeOpen
Prompt AI EditzOpen
Ronytica alternativeOpen
Pixlr alternativeOpen
Free AI image generator, no restrictionsOpen
MiniMax H3 Image character swapOpen
MiniMax H3 Image ComfyUI workflowsOpen
SeedanceOpen
LTX 2.5Open
PAN2Open
FastFrames LOpen
Swagga AI websiteOpen
Wan 2.2 and Wan 2.1Open
M4 Editor alternativeOpen
Harish Edits alternativeOpen- AI photo effectsOpen
- All AI photo effectsOpen
FAQ
Questions, answered
What is LongCat Video Avatar 1.6?
It is an open-source model from Meituan’s LongCat Team for audio-driven human video: speech goes in, with an optional reference photo, and a talking person comes out. It upgrades Avatar 1.5, is built on the LongCat-Video base model, and its weights are MIT-licensed.
Is this site LongCat Video Avatar 1.6, and can I make an avatar video here?
No. This site is not LongCat or Meituan and does not run Avatar 1.6. It currently delivers images only, not video. Its own editor uses a different model: upload one photo and write the change (up to 400 characters), or tap a ready-made edit.
How do I run LongCat Video Avatar 1.6 from GitHub, and does it work in ComfyUI?
The steps are in the GitHub README of meituan-longcat/LongCat-Video: a Python 3.10 environment, torch 2.6.0, flash_attn, two requirements files, weights from Hugging Face, then a torchrun demo. We found no confirmed ComfyUI route for 1.6; a community news site said the existing community pack had not been updated for it.
How much GPU memory does LongCat Video Avatar 1.6 need?
Not stated. Neither the model card nor the README gives a memory minimum. What we can give is download size, summed by us from the file list on 2026-11-19: 31.71 GB for the BF16 base model and 15.88 GB for the optional INT8 version, plus 2.52 GB per adapter.
Can I use it commercially? What is the licence?
The weights are MIT, and the card says no rights to Meituan trademarks or patents are granted. But the project page says generated content is for academic use only and commercial use is not permitted. We could not reconcile the two, so ask Meituan before any commercial use. This is not legal advice.
What changed from LongCat Avatar 1.5 to 1.6?
Per the card, 1.6 keeps the 1.5 architecture and adds two LoRA adapters. The makers claim more natural lip movement, steadier clips past 10 minutes and faster inference, and their own preference study puts 1.6 ahead of 1.5. We have not tested any of it.
How is it different from LongCat Video?
LongCat Video is the base video model (text-to-video, image-to-video, continuation); Avatar 1.6 is a separate audio-driven model built on it. Our LongCat Video page covers the base model, its MIT licence and fal.ai’s per-second prices.
Can I try something here for free?
Partly: new accounts get 10 free credits after signing in, enough for 1 free image (once per account, valid for 7 days, while the day's new-account gifts last). Each image costs 10 credits, from $0.13 per image on a monthly plan.
Just need one photo edited?
Upload it, write the change, and download the result in about 37 seconds.
Edit a photo hereIndependent site. Not affiliated with Meituan, LongCat, Hugging Face, ComfyUI or fal.ai, and this site does not run LongCat Video Avatar 1.6 or make video. Facts about the model come from the makers’ own pages and the third-party pages named above, read on 2026-11-19; the makers’ claims are not tested by us. This is not legal advice.