Cheap Gemini 3 Flash API
The Flash Gemini tier is all about speed and price — ideal for high-volume work: chat, classification, parsing images and documents by the thousand. Through cheapai it's available via an OpenAI-compatible API from $0.12/1M per 1M tokens, roughly 77% cheaper than Google's official price. No VPN, pay in ₽, USDT or by card, and it works out of the box with Cursor, Claude Code, Cline and any OpenAI SDK.
| Input / 1M | $0.12 |
| Output / 1M | $0.70 |
| Cache read / 1M | $0.01 |
What it's good for
- high-volume chat and classification
- image and document processing
- low-cost agentic pipelines
How to connect Gemini 3 Flash
- 1. Register and top up: cheapai.io/register
- 2. Create an API key: cheapai.io/keys
- 3. Point your tool to our API:
Base URL: https://cheapai.io/v1 API key: sk-... (from /keys) Model: gemini-3-flash
Works with Cursor, Claude Code, Cline, Codex and any OpenAI-compatible SDK. Setup guides →
FAQ
How much does Gemini 3 Flash cost via cheapai?
$0.12/1M per 1M tokens for input, $0.70/1M for output — about 77% below Google's official price.
How do I connect Gemini 3 Flash?
Point your tool's base URL to https://cheapai.io/v1, create an API key at cheapai.io/keys, and use model "gemini-3-flash". It works with Cursor, Claude Code, Cline and any OpenAI-compatible SDK.
Gemini 3 Flash API pricing — what exactly am I charged for?
Tokens, nothing else: $0.12 per 1M tokens you send in, $0.70 per 1M the model writes back, and $0.01 per 1M read from prompt cache. No subscription, no per-request fee, and the balance doesn't expire.
Is Gemini 3 Flash cheaper here than buying it direct?
Yes — about 77% cheaper on input than Google charges, and the price above is what you actually pay, with no minimum monthly spend.
Do I need a VPN?
No — cheapai works without a VPN and takes ₽, USDT and cards.