GEMMA API
Google Gemma, hosted and ready.
Call Google's open-weight Gemma 3 through an OpenAI-compatible endpoint: strong instruction following in a 27B model, with 128K context.
Hand this to your agent
Add a Runcrate model to this project: run curl -fsSL"https://runcrate.ai/agents.md?intent=build"and follow it.
Your agent asks for an API key. Create a key
Quick start
Hand it to your agent. Or read the request it sends.
Your agent writes and runs it. The same key works for every model and every GPU.
Available models
Models you can call today. Billed prices, as charged.
| Model | Maker | Price | Detail |
|---|---|---|---|
| google/gemma-3-27b-it | $0.084 in, $0.168 out per 1M tokens | 27B, instruction-tuned, 128K context | |
| google/gemma-3-4b-it | $0.0525 in, $0.105 out per 1M tokens | 4B, lightweight, fast |
Why Runcrate
Built for production. One key, one balance.
Gemini heritage
Built on the same research as Google's Gemini. Gemma distills frontier-model quality into open weights you can run anywhere.
Compact and efficient
At 27B parameters, Gemma 3 punches above its weight class. Strong reasoning and instruction following at lower inference cost than 70B models.
128K context
Long context window for document analysis, codebase comprehension, and extended conversations without truncation.
Open weights
Use the API for convenience, or download the weights and self-host on Runcrate GPU instances for full control over serving configuration.
FAQ