MedGemma models
4 MedGemma models from Google on Hugging Face, from 4.3B to 28.8B parameters, published by Google in BF16, with a 128K-token context. The smallest official model, medgemma-1.5-4b-it, needs about 9.6 GB of VRAM at its published precision; the cheapest live fit is RTX 3060 at $0.110/hr.
Part of the Gemma series
Pick a size
One row per official MedGemma size: the VRAM it needs at each precision, the cheapest GPU that holds it today, and the KV cache for a 32K-token context.
| Model | Parameters | Native VRAM | FP8 VRAM | INT4 VRAM | Live GPU fit (native) | Est. $/hr | KV cache at 32K |
|---|---|---|---|---|---|---|---|
| medgemma-4b-it | 4.3B | 9.6 GB | 4.8 GB | 2.4 GB | RTX 3060 | $0.110/hr | 0.74 GB |
| medgemma-27b-text-it | 27.0B | 60.4 GB | 30.2 GB | 15.1 GB | A100 | $1.31/hr | 2.9 GB |
| medgemma-27b-it | 28.8B | 64.5 GB | 32.2 GB | 16.1 GB | A100 | $1.31/hr | 2.9 GB |
VRAM is weights times a flat 1.2 overhead; the KV cache is a separate per-model figure at 16-bit, shown where the architecture is published. See the methodology.
Official models (4)
| Model | Parameters | Published as | VRAM needed | Live GPU fit | Est. $/hr |
|---|---|---|---|---|---|
| medgemma-1.5-4b-it | 4.3B | BF16 | 9.6 GB | RTX 3060 | $0.110/hr |
| medgemma-4b-it | 4.3B | BF16 | 9.6 GB | RTX 3060 | $0.110/hr |
| medgemma-27b-text-it | 27.0B | BF16 | 60.4 GB | A100 | $1.31/hr |
| medgemma-27b-it | 28.8B | BF16 | 64.5 GB | A100 | $1.31/hr |
VRAM is for the precision the model is published in: the weights times a flat 1.2 overhead, with the KV cache not included. 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.