ByteDance-Seed models
3 ByteDance-Seed models on Hugging Face, from 8.3B to 36.2B parameters. At the precision each one is published in, the smallest needs about 20.9 GB of VRAM (academic-ds-9B, cheapest live fit: RTX A5000) and the largest about 80.8 GB (Seed-OSS-36B-Instruct, cheapest live fit: RTX PRO 6000). The cheapest way to run academic-ds-9B is $0.176/hr.
ByteDance-Seed models
| Model | Parameters | Published as | VRAM needed | Live GPU fit | Est. $/hr |
|---|---|---|---|---|---|
| UI-TARS-1.5-7B | 8.3B | F32 | 37.1 GB | RTX A6000 | $0.363/hr |
| academic-ds-9B | 9.4B | BF16 | 20.9 GB | RTX A5000 | $0.176/hr |
| Seed-OSS-36B-Instruct | 36.2B | BF16 | 80.8 GB | RTX PRO 6000 | $1.38/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.
