LLM Opex

Models

Google models: prices and the hardware each needs

26 models, 5 sold per token by 7 vendors. Vendors are in alphabetical order; the price is the cheapest list price among them, per million tokens.

ModelSizeSold per token byFrom, inputOutputGPU memory
codegemma-7b-it8.54 Bnot sold per token21.3 GB
Gemini 2.5 FlashclosedDeepInfra, Gemini API, Snowflake$0.30$2.50
Gemini 2.5 Flash LiteclosedDatabricks, Gemini API, Snowflake$0.10$0.40
Gemini 2.5 ProclosedDeepInfra, Gemini API$1.25$10.00
Gemma 3 27B27.4 BDeepInfra, Nebius$0.08$0.1668.6 GB
Gemma 4 26B A4B25.8 BCloudflare, DeepInfra, Snowflake, Telnyx$0.07$0.3464.5 GB
gemma-1.1-2b-it2.51 Bnot sold per token6.3 GB
gemma-2-27b-it27.23 Bnot sold per token68.1 GB
gemma-2-2b2.61 Bnot sold per token13 GB
gemma-2-2b-it (google)2.61 Bnot sold per token6.5 GB
gemma-2-9b9.24 Bnot sold per token46.2 GB
gemma-2-9b-it9.24 Bnot sold per token23.1 GB
gemma-2b2.51 Bnot sold per token6.3 GB
gemma-2b-it (google)2.51 Bnot sold per token6.3 GB
gemma-3-12b-it12.19 Bnot sold per token30.5 GB
gemma-3-1b-it (google)1 Bnot sold per token2.5 GB
gemma-3-1b-pt1 Bnot sold per token2.5 GB
gemma-3-270m-it0.27 Bnot sold per token0.7 GB
gemma-3-4b-it4.3 Bnot sold per token10.8 GB
gemma-4-12B-it11.96 Bnot sold per token29.9 GB
gemma-4-26B-A4B26.54 Bnot sold per token66.3 GB
gemma-4-31b-it31.27 Bnot sold per token78.2 GB
gemma-7b8.54 Bnot sold per token21.3 GB
gemma-7b-it8.54 Bnot sold per token21.3 GB
medgemma-27b-text-it27.01 Bnot sold per token67.5 GB
shieldgemma-9b9.24 Bnot sold per token23.1 GB