The whole Gemini lineup
A very large context window
Gemini’s traditional strong suit is its huge context: whole reports, long logs and large codebases fit in the window without chunking. That’s handy for document parsing, RAG with large fragments, and analyzing long conversations. The exact limit depends on the version and matches Google’s official one — check the current values on the Pricing page at www.ruapi.ai.Multimodal: reads images and generates them
Gemini works with images both ways. On input (vision) the models take images alongside text — see the image input guide for how to send an image in the request. On output dedicated models generate and edit images:
gemini-nano-banana-2.1, gemini-3-pro-image, gemini-3.1-flash-image, gemini-3.1-flash-lite-image and others — see image generation. The exact current model names are on the Pricing page.gemini-4-argon, gemini-3.8-flash, gemini-3.7-flash, gemini-3.6-flash, gemini-3.5-flash, gemini-3.5-flash-lite, gemini-3.1-flash-lite, gemini-3.1-pro-preview, gemini-3-flash-preview, gemini-2.5-pro, gemini-2.5-flash, gemini-2.5-flash-lite. They can call functions and stream the response token by token.
Which version to pick (2.5 / 3 / 3.1 / 3.5 / 3.6–3.8 / 4)
- 2.5 (
gemini-2.5-pro,gemini-2.5-flash,gemini-2.5-flash-lite) — the mature, stable generation and a good default choice. - 3 / 3.1 — newer models, including preview versions (
gemini-3-flash-preview,gemini-3.1-pro-preview) and image generation; reach for them for the latest capabilities, keeping in mind preview names can change. The stablegemini-3.1-flash-litecosts less thangemini-3.5-flash-lite. - 3.5 (
gemini-3.5-flash,gemini-3.5-flash-lite) — the fresh Flash generation:-flash-liteis noticeably cheaper and faster,-flashis stronger. - 3.6 / 3.7 / 3.8 (
gemini-3.6-flash,gemini-3.7-flash,gemini-3.8-flash) — the newest Flash generations, all at the same price;gemini-3.8-flashis the latest and the current pick for new code. - 4 (
gemini-4-argon) — Google’s flagship reasoning model, released in October 2026. About 1.05M tokens of context (input and output combined), up to 65,536 output tokens per response, and it understands images. The current price is Google’s limited-time offer; from 1 December 2026 the price doubles — see the changelog.
Preview models change names as newer versions ship. Don’t hardcode IDs — the exact current names and per-token prices are always on the Pricing page at www.ruapi.ai.
Connecting Gemini
You don’t need Google’s separate library: RuAPI accepts requests in the OpenAI format, so take any OpenAI SDK and setbase_url to the address below.
- base_url:
https://www.ruapi.ai/v1 - model: e.g.
gemini-3.8-flash
A page for each model
Specs from Google’s official documentation, the current RuAPI price and an FAQ — one page per model:- Gemini 4 Argon —
gemini-4-argon - Gemini 3.1 Pro —
gemini-3.1-pro-preview - Gemini 3.5 Flash-Lite —
gemini-3.5-flash-lite - Nano Banana 2.1 —
gemini-nano-banana-2.1 - Nano Banana Pro —
gemini-3-pro-image(nano-banana-pro)
Next
Connect and first request
base_url, key, Python and curl code.
Claude models
Same key — Claude’s lineup and capabilities.
GPT models
OpenAI’s models — same address, same key.