Video models · Rhymes AI
Allegro (2.8B)
- Backbone
- 2.8B
- Native output
- 1280×720, 88 frames
- Default steps
- 100
- License
- apache-2.0
- Released
- 2024-10
How to run it
Original weights · Hugging Face
Full-precision files from Rhymes AI, for fine-tuning, vLLM, Transformers, Diffusers or ComfyUI.
Official repo · GitHub
Reference code, inference scripts and docs from Rhymes AI. The weights themselves download from Hugging Face.
github.com/rhymes-ai/Allegro (opens in a new tab)
git clone https://github.com/rhymes-ai/AllegroMemory needed
Peak GPU memory at the native output size. Offloading moves the text encoder, or the whole pipeline, to system RAM between steps.
| Precision | Everything on GPU | Text encoder offloaded | Sequential offload | Download |
|---|---|---|---|---|
FP16 | 38.2 GB | 28.7 GB | 24.7 GB | Original weights (opens in a new tab) |
Components
- Backbone: 2.8B
- VAE: 0.17B
- T5-XXL: 4.76B
Best hardware for Allegro (2.8B)
The least expensive GPU in our catalog that runs it without offloading is the Intel Arc Pro B70 (launch price $949) at FP16.
| Quality | Hardware | Quant | Memory | Speed | Runs |
|---|---|---|---|---|---|
| 52Fair quality (C) · 52 of 100: how capable the model is at FP16, whatever it runs on. Match score 64/100 (quality, fit, speed and memory use) sets the ranking. |  NVIDIA RTX PRO 6000 Blackwell |
FP16 | 38.2 GB / 96.0 GB | 14.7 min | Slow |
| 52Fair quality (C) · 52 of 100: how capable the model is at FP16, whatever it runs on. Match score 62/100 (quality, fit, speed and memory use) sets the ranking. |  NVIDIA GeForce RTX 5090 |
FP16 | 28.7 GB / 32.0 GB | 17.6 min | Slowtight fit · encoder in RAM |
| 52Fair quality (C) · 52 of 100: how capable the model is at FP16, whatever it runs on. Match score 61/100 (quality, fit, speed and memory use) sets the ranking. |  AMD Ryzen AI Max+ 395 (128 GB) |
FP16 | 38.2 GB / 96.0 GB | 38.4 min | Slow |
| 52Fair quality (C) · 52 of 100: how capable the model is at FP16, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking. |  NVIDIA L40S |
FP16 | 38.2 GB / 48.0 GB | 18.5 min | Slow |
| 52Fair quality (C) · 52 of 100: how capable the model is at FP16, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking. |  NVIDIA RTX 6000 Ada Generation |
FP16 | 38.2 GB / 48.0 GB | 18.6 min | Slow |
| 52Fair quality (C) · 52 of 100: how capable the model is at FP16, whatever it runs on. Match score 69/100 (quality, fit, speed and memory use) sets the ranking. |  NVIDIA L40 |
FP16 | 38.2 GB / 48.0 GB | 18.7 min | Slow |
| 52Fair quality (C) · 52 of 100: how capable the model is at FP16, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking. |  AMD Radeon PRO W7900 |
FP16 | 38.2 GB / 48.0 GB | 22.3 min | Slow |
| 52Fair quality (C) · 52 of 100: how capable the model is at FP16, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking. |  NVIDIA RTX PRO 5000 Blackwell 48GB |
FP16 | 38.2 GB / 48.0 GB | 23.9 min | Slow |
| 52Fair quality (C) · 52 of 100: how capable the model is at FP16, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking. |  NVIDIA L20 |
FP16 | 38.2 GB / 48.0 GB | 25.5 min | Slow |
| 52Fair quality (C) · 52 of 100: how capable the model is at FP16, whatever it runs on. Match score 68/100 (quality, fit, speed and memory use) sets the ranking. |  AMD Radeon PRO W7800 48GB |
FP16 | 38.2 GB / 48.0 GB | 28.0 min | Slow |
Estimates come from published hardware specs and our open formulas, not from running the model on your machine. Real speed varies with runtime, drivers and settings. Numbers marked “Measured” come from a linked benchmark.