Gemma 2 models

11 Gemma 2 models from Google on Hugging Face, from 2.5B to 27.2B parameters, published by Google in BF16 and F32, with a 8K-token context. The smallest official model, gemma-2b, needs about 5.6 GB of VRAM at its published precision; the cheapest live fit is RTX 3070 at $0.088/hr.

Part of the Gemma series · Previous generation: Gemma 1 · Next generation: Gemma 3

Pick a size

One row per official Gemma 2 size: the VRAM it needs at each precision, the cheapest GPU that holds it today, and the KV cache for a 32K-token context.

ModelParametersNative VRAMFP8 VRAMINT4 VRAMLive GPU fit (native)Est. $/hrKV cache at 32K
gemma-2b-it2.5B5.6 GB2.8 GB1.4 GBRTX 3070$0.088/hr0.56 GB
gemma-2-2b-it2.6B5.8 GB2.9 GB1.5 GBRTX 3070$0.088/hr1.8 GB
gemma-2-9b-it9.2B20.7 GB10.3 GB5.2 GBRTX A5000$0.176/hr5.9 GB
gemma-2-27b-it27.2B60.9 GB30.4 GB15.2 GBA100$1.31/hr6.5 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 (8)

ModelParametersPublished asVRAM neededLive GPU fitEst. $/hr
gemma-2b2.5BBF165.6 GBRTX 3070$0.088/hr
gemma-2b-it2.5BBF165.6 GBRTX 3070$0.088/hr
gemma-2-2b2.6BF3211.7 GBV100$0.110/hr
gemma-2-2b-it2.6BBF165.8 GBRTX 3070$0.088/hr
gemma-2-9b9.2BF3241.3 GBRTX A6000$0.363/hr
gemma-2-9b-it9.2BBF1620.7 GBRTX A5000$0.176/hr
gemma-2-27b27.2BF32122 GBAMD MI300X$2.63/hr
gemma-2-27b-it27.2BBF1660.9 GBA100$1.31/hr

Fine-tunes and community models (3)

ModelParametersPublished asVRAM neededLive GPU fitEst. $/hr
gemma-2-2b2.6BBF165.8 GBRTX 3070$0.088/hr
gemma-2-9b9.2BBF1620.7 GBRTX A5000$0.176/hr
gemma-2-27b27.2BBF1660.9 GBA100$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.

Run and fine-tune Gemma 2

More from Google

Start ComfyUI on a rented GPU.

© 2026 Aquanode. All rights reserved.

All trademarks, logos and brand names are the property of their respective owners.