Cosmos models
6 Cosmos models on Hugging Face, from 2.4B to 64.6B parameters. At the precision each one is published in, the smallest needs about 5.5 GB of VRAM (Cosmos-Reason2-2B, cheapest live fit: RTX 5060 Ti) and the largest about 144 GB (Cosmos3-Super-Image2Video, cheapest live fit: RTX A5000). The cheapest way to run Cosmos-Reason2-2B is $0.110/hr.
Cosmos models
| Model | Parameters | Published as | VRAM needed | Live GPU fit | Est. $/hr |
|---|---|---|---|---|---|
| Cosmos-Reason2-2B | 2.4B | BF16 | 5.5 GB | RTX 5060 Ti | $0.110/hr |
| Cosmos-Reason1-7B | 8.3B | BF16 | 18.5 GB | RTX A5000 | $0.176/hr |
| Cosmos-Reason2-8B | 8.8B | BF16 | 19.6 GB | RTX A5000 | $0.176/hr |
| Cosmos-H-Surgical | 15.2B | BF16 | 33.9 GB | RTX A6000 | $0.363/hr |
| Cosmos3-Super-Text2Image-4Step | 64.0B | BF16 | 143 GB | RTX A5000 × 6 | $1.06/hr |
| Cosmos3-Super-Image2Video | 64.6B | BF16 | 144 GB | RTX A5000 × 7 | $1.23/hr |
VRAM is for the precision the model is published in, with the same overhead and an 8,192-token context assumed on every page; see the methodology. The fit shown is the lowest-priced single GPU type that holds the model at that precision, or the lowest-priced multi-GPU set (up to 8) when none does.
More from NVIDIA
- Parakeet (7 models)
- Nemotron 3 (16 models)
- Nemotron Nano (4 models)
- Nemotron (5 models)
- Nemotron Labs (4 models)
- Nemotron-H (3 models)
