Gemma 1 models
3 Gemma 1 models from Google on Hugging Face, from 2.5B to 8.5B parameters, published by Google in BF16, with a 8K-token context. The smallest official model, gemma-1.1-2b-it, needs about 5.6 GB of VRAM at its published precision; the cheapest live fit is RTX 4070 Super at $0.121/hr.
Part of the Gemma series · Next generation: Gemma 2
Pick a size
One row per official Gemma 1 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 |
|---|---|---|---|---|---|---|---|
| gemma-1.1-2b-it | 2.5B | 5.6 GB | 2.8 GB | 1.4 GB | RTX 4070 Super | $0.121/hr | 0.56 GB |
| gemma-7b | 8.5B | 19.1 GB | 9.5 GB | 4.8 GB | RTX A5000 | $0.176/hr | 14.0 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 (3)
| Model | Parameters | Published as | VRAM needed | Live GPU fit | Est. $/hr |
|---|---|---|---|---|---|
| gemma-1.1-2b-it | 2.5B | BF16 | 5.6 GB | RTX 4070 Super | $0.121/hr |
| gemma-7b | 8.5B | BF16 | 19.1 GB | RTX A5000 | $0.176/hr |
| gemma-7b-it | 8.5B | BF16 | 19.1 GB | RTX A5000 | $0.176/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.
