Phi-3 models
6 Phi-3 models on Hugging Face, from 3.8B to 14.0B parameters. At the precision each one is published in, the smallest needs about 8.5 GB of VRAM (Phi-3-mini-128k-instruct, cheapest live fit: RTX 5060 Ti) and the largest about 31.2 GB (Phi-3-medium-128k-instruct, cheapest live fit: RTX 4080 Super). The cheapest way to run Phi-3-mini-128k-instruct is $0.110/hr.
Phi-3 models
| Model | Parameters | Published as | VRAM needed | Live GPU fit | Est. $/hr |
|---|---|---|---|---|---|
| Phi-3-mini-128k-instruct | 3.8B | BF16 | 8.5 GB | RTX 5060 Ti | $0.110/hr |
| Phi-3-mini-4k-instruct | 3.8B | BF16 | 8.5 GB | RTX 5060 Ti | $0.110/hr |
| Phi-3-vision-128k-instruct | 4.1B | BF16 | 9.3 GB | RTX 5060 Ti | $0.110/hr |
| Phi-3-small-8k-instruct | 7.4B | BF16 | 16.5 GB | RTX A5000 | $0.176/hr |
| Phi-3-medium-128k-instruct | 14.0B | BF16 | 31.2 GB | RTX 4080 Super | $0.338/hr |
| Phi-3-medium-4k-instruct | 14.0B | BF16 | 31.2 GB | RTX 4080 Super | $0.338/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 Microsoft
- Florence-2 (5 models)
- Phi-4 (7 models)
- Phi-3.5 (4 models)
- VibeVoice (4 models)
- microsoft (6 models)
