Model profile

FramePack AI video generator

An open long-video/image-to-video technique that compresses frame context to extend generation length.

Home/Models/FramePack

Decision Snapshot

Score: 51

Score source: Provisional quality proxy based on official/research evidence and catalogue presence; exact external rank requires live leaderboard/API collection.. Evidence confidence is medium.

Best For

  • Long-Video Experiments
  • Image-To-Video Research
  • Open Workflow Testing

Supported Workflows

  • I2v
  • Long Video
  • Open Weights

Access And Company

Built by Lllyasviel. Public access is listed as Open Weights, Research Only.

Entry Points

Score Breakdown

Weighted evidence dimensions separate model quality from access, web authority, ecosystem, and source confidence.

DimensionScoreWeightEvidenceConfidence
Model Quality Proxy 60 24% Provisional quality proxy based on official/research evidence and catalogue presence; exact external rank requires live leaderboard/API collection. low
Missing: live Artificial Analysis rank/Elo by modality, live Arena AI rank/score by modality
Capability Depth 56 16% Supported workflow tags: i2v, long_video, open_weights. Depth rewards T2V/I2V/V2V, references, native audio, editing, API, and open weights where present. medium
Missing: hands-on feature verification, mode-specific limits by duration/resolution
Access And Pricing 30 12% Access types: open_weights, research_only. Pricing note: Open/self-hosted; cost depends on hardware and integrated model. low
Missing: current free tier, normalized price per video minute
Version Maturity 28 12% 1 version/history entries in the profile; latest public date is 2025. medium
Missing: automated release-note monitor, model ID/version mapping across providers
Ecosystem Popularity 100 10% GitHub snapshot includes 1 repo(s), 17,161 stars, 1,731 forks, 486 open issues, latest pushed 2025-10-16T01:28:50Z. high
Missing: Hugging Face downloads/likes, Replicate/fal usage/runs
Web Authority 28 10% SEO snapshot found product domain github.com and company domain unknown; sampled sitemap URL count is unknown.

No public source link is attached yet.

 medium
Missing: Similarweb/API traffic, Tranco rank
Company Distribution 30 10% Company profile is missing; distribution score is provisional. low
Missing: app store ratings/review volume, product MAU/traffic
Source Confidence 74 6% Profile source confidence is partial with 1 official URL(s), 1 feature evidence item(s), SEO present, GitHub present. medium
Missing: paid/API evidence refresh, manual output review artifacts

What It Is

Primary content is rendered into static HTML.

FramePack is an open community project from lllyasviel focused on longer video generation using frame-context packing. It is a method and workflow direction more than a polished consumer model.

For the benchmark site, include FramePack under open/research because long-video continuity is one of the most important evaluation dimensions for future AI video tools.

Key Features

  • Frame-context packingFramePack targets longer video generation by packing frame context efficiently.Source
  • Primary source-backed profileFramePack has an official source, repository, paper, or model card attached for verification.Source
  • Open/self-hosted access pathFramePack can be evaluated through open code, weights, or model-card resources when license terms allow.Source

Version Progress

  • 2025 / partial / partialFramePackCommunity release focused on longer video generation workflows.

Latest News

Model-specific news entries link directly to the original publisher; article bodies stay off-site.

Limitations And Pricing

Open/self-hosted; cost depends on hardware and integrated model.

  • Better Treated As A Method/Workflow Than A Standalone Commercial Model.
  • Needs Active Maintenance Review.

Sources