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 720x1280 and 15 fps, roughly 6 seconds, with optional interpolation to 30 fps.
- Up to
- 6 seconds
- Resolution
- 720x1280
- Weights
- Downloadable
- API
- No
- Commercial use
- Vendor states commercial use is permitted
Its distinguishing feature was fully permissive licensing at a small parameter count.
Apache-2.0 for both weights and code, with no revenue threshold, MAU cap, territorial restriction or acceptable-use addendum.
Specification
- What it takes in
- text
- What it makes
- video
- Maximum duration
- 6 seconds
- Maximum resolution
- 720x1280
- Aspect ratios
- not established
- Controls
- text prompt, negative prompt, guidance scale, inference steps, seed, CPU offload, frame interpolation to 30 fps (EMA-VFI)
- Weights
- Downloadable
- API
- No
- Pricing
- Free open weights (Apache-2.0), self-hosting only with no first-party API. Documented at 9.3GB VRAM in BF16 with CPU offloading, which makes it comfortably consumer-GPU viable.
- Released or updated
- 2024-10-20
Supported aspect ratios is not established here. That means no page we could retrieve stated it, not that the model lacks one.
Strengths
- Apache-2.0 with no restrictions
- 720x1280 output from a 2.8B model
- 9.3GB VRAM with offloading
- Documented frame count and fps so duration is verifiable
- Official interpolation path to 30 fps
Weaknesses
- 15 fps native output needs interpolation to look smooth
- Quality is far behind 2025-2026 models
- Text-to-video only; no image conditioning in the base release
- Effectively unmaintained since late 2024
- Very small community and tooling footprint
Best for
- Consumer-GPU experimentation and cases where Apache-2.0 licensing outweighs output quality.
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
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
ByteDance Seedance 2.0
Seedance 2.0 is ByteDance Seed's unified multimodal audio-video model: a single generation pass accepts text plus mixed references (up to 9 images, 3
Video generation