Shotcraft
Video generation

Open-Sora 2.0

An 11B-parameter open video generation model released with weights, inference code and training code under Apache-2.0. It generates at 256px and 768px in frame counts expressed as 4k+1 up to 129 frames.

A video generation model by HPC-AI Tech (Colossal-AI) · read 2026-08-14 · sourcing high

Resolution
768px
Weights
Downloadable
API
No
Commercial use
Vendor states commercial use is permitted

HPC-AI Tech published its approximately $200K training cost, making it the most transparent open video project on training economics.

Commercial use — Vendor states commercial use is permitted

Apache-2.0 for both code and weights, with no revenue threshold, MAU cap, territorial restriction or acceptable-use addendum.

The licence notes recorded during research were read for a commercial-use position and did not state one. Read the licence before you rely on any of this. Nothing on this page is legal advice, and this site does not tell you a model is cleared for commercial use.

Specification

Vendor figures, read 2026-08-14Read from HPC-AI Tech (Colossal-AI)'s own documentation on 2026-08-14. Vendors change these between model versions without announcing it, so check theirs before you commit a shot to a number. Sourcing for this entry is strong: the figures came from pages we retrieved and read. Check the vendor's current documentation

What it takes in
text, image
What it makes
video
Maximum duration
not established
Maximum resolution
768px
Aspect ratios
not established
Controls
text prompt, reference image, resolution selection (256px/768px), frame count in 4k+1 steps, motion score, aesthetic score conditioning, multi-GPU sequence parallelism, full training pipeline
Weights
Downloadable
API
No
Pricing
Free open weights (Apache-2.0), self-hosting only - there is no first-party hosted API. VRAM is high: the repository documents 52.5GB on a single GPU at 256x256 falling to 44.3GB with multi-GPU, and 60.3GB single-GPU at 768x768 falling to 44.3GB across 8 GPUs. Practically a datacenter-GPU model.
Released or updated
2025-03-12

Maximum duration and supported aspect ratios are not established here. That means no page we could retrieve stated them, not that the model lacks them.

Strengths

  • Apache-2.0 across weights, inference and training code
  • Published training cost (~$200K) and full reproduction recipe
  • Claimed parity with 11B HunyuanVideo and 30B Step-Video at far lower training cost
  • Best choice for anyone who needs to actually retrain rather than fine-tune

Weaknesses

  • 44-60GB VRAM even at low resolution
  • 768px ceiling is low by 2026 standards
  • No fps documented, so clip length is unclear
  • No first-party hosted API
  • Development cadence has slowed relative to Wan and Hunyuan

Best for

What to understand before using it

Explainers naming the mechanisms this model's own documentation mentions.

See it from the vendor

Shotcraft hosts no sample output. Judging a model means watching its own showreel with your own eyes, and a clip re-hosted here would be a clip we chose.

Nearby in the directory

Adobe Firefly Video Model

The Adobe Firefly Video Model generates video from a text prompt, optionally guided by an image used as the first or final frame. Adobe's Firefly Serv

Video generation

Allegro

A small text-to-video model comprising a 2.8B VideoDiT and a 175M VideoVAE, released under Apache-2.0 in October 2024. It generates 88-frame clips at

Video generation

Amazon Nova Reel 1.1

Amazon Nova Reel generates video from text prompts or from text plus an image, and is available only through Amazon Bedrock. AWS documents output at 1

Video generation

Sources

Read on 2026-08-14. A source that has changed since then will not be reflected here until the next pass.