Gemma API pricing: Gemma 4 31B, Gemma 4 26B A4B and Gemma 3 at every provider

Updated Nov 1, 2026

Gemma is Google's family of open-weight models. Google sells some of them on Vertex AI, but because the weights are open, most Gemma offers come from inference clouds and API resellers - over thirty providers sell Gemma 4 31B alone. Below: the typical price (the median across providers) of the Gemma models sold most widely, and live tables of every seller. The tables are rebuilt from today's prices on every page view.

The models in one list

Price per 1M tokens, input / output, in the 31 October 2026 collection; "typical" is the median across providers:

• Gemma 4 31B - typically $0.14 / $0.40, input from $0.09; up to 262K context, 33 providers. The largest Gemma.
• Gemma 4 26B A4B - typically $0.11 / $0.40, input from $0.05; up to 262K context, 22 providers. A mixture-of-experts model with about 4B active parameters - fast and cheap to serve.
• Gemma 3 27B - typically $0.12 / $0.22; 131K context, 13 providers.
• Gemma 3 12B - typically $0.05 / $0.15, 12 providers.
• Gemma 3 4B - typically $0.05 / $0.10, 8 providers. For simple, high-volume jobs.

Gemma 4 31B: every provider

Live data: this table is built from today's prices.

Gemma 4 26B A4B: every provider

Live data: this table is built from today's prices.

Gemma 3 27B: every provider

Live data: this table is built from today's prices.

How to choose

• The strongest Gemma: Gemma 4 31B - the spread between providers is wide, so compare before you pick.
• Speed and a low price: Gemma 4 26B A4B - output costs the same as 31B at the median, input is lower.
• Volume at the lowest price: Gemma 3 12B or 4B.
• Other cheap open models: the cheapest LLM APIs, Qwen, Llama.
• Every Gemma offer in one list: Gemma in the catalog.

All prices are in USD per 1 million tokens, without taxes, as the providers publish them; each row links to its card with the source and the date it was checked.