Google · 28 models · multi-supplier bidding · get a key
How do I use the Gemini API and what does it cost? Create a key in the console: OpenAI protocol sets base_url to https://321ai.cc/v1; Anthropic protocol (Claude Code) sets ANTHROPIC_BASE_URL to https://321ai.cc. This page lists 28 Gemini models with live market floor prices from competing suppliers (data date 2026-11-20; machine-readable: https://321ai.cc/api/public/model-prices). Failures never billed, platform fee 0.
Google Gemini family — long context and multimodal strengths.
| Model | Market floor price | Channels |
|---|---|---|
| gemini-4-argon | $0.1305 in · $0.6525 out / 1M tok | 338 |
| gemini-3.8-flash | $0.01296 in · $0.0648 out / 1M tok | 326 |
| gemini-4-hydrogen | $0.04302 in · $0.2151 out / 1M tok | 77 |
| gemini-4-carbon | $0.015 in · $0.075 out / 1M tok | 67 |
| gemini-3.7-flash | $0.013886 in · $0.069428 out / 1M tok | 58 |
| gemini-3.6-flash | $0.015 in · $0.075 out / 1M tok | 39 |
| gemini-3.1-pro-preview | $0.015 in · $0.09 out / 1M tok | 36 |
| gemini-3-flash-preview | $0.005988 in · $0.035928 out / 1M tok | 32 |
| gemini-3.1-flash-image | $0.00666 / call | 29 |
| nano-banana-pro | $0.0108 / call | 25 |
| gemini-2.5-flash | $0.0072 in · $0.06 out / 1M tok | 24 |
| gemini-3.1-flash-lite-preview | $0.018 in · $0.108 out / 1M tok | 23 |
| gemini-3.5-flash | $0.0162 in · $0.0972 out / 1M tok | 23 |
| gemini-3-flash-agent | $0.015 in · $0.09 out / 1M tok | 20 |
| nano-banana-2 | $0.006657 / call | 20 |
| gemini-2.5-pro | $0.015 in · $0.12 out / 1M tok | 17 |
| gemini-3.5-flash-lite | $0.018 in · $0.15 out / 1M tok | 16 |
| nano-banana-2.1 | $0.0081 / call | 15 |
| gemini-3-pro-image-preview | $0.01275 / call | 14 |
| gemini-3-pro-preview | $0.015 in · $0.09 out / 1M tok | 13 |
| gemini-3.5-flash-extra-low | $0.015 in · $0.09 out / 1M tok | 10 |
| gemini-3.5-flash-high | $0.015 in · $0.09 out / 1M tok | 9 |
| gemini-3.1-pro-preview-low | $0.0936 in · $0.5616 out / 1M tok | 8 |
| gemini-omni-1.1-flash | $0.04104 / call | 8 |
| nano-banana | $0.0054 / call | 8 |
| gemini-embedding-2-preview | $0.2097 in · <$0.0001 out / 1M tok | 6 |
| gemini-3-pro-image | $0.0294 / call | 3 |
| gemini-3.1-flash-image-preview | $1.500 / call | 2 |
Floor across suppliers; per-supplier quotes on the marketplace. Every model page has a .md machine-readable twin.
Set base_url to https://321ai.cc/v1; pick any model above as model.
from openai import OpenAI
client = OpenAI(api_key="YOUR_KEY", base_url="https://321ai.cc/v1")
resp = client.chat.completions.create(
model="gemini-4-argon",
messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)
All 28 Gemini models are quoted by multiple suppliers; this page shows the market floor price per model, refreshed live (data date 2026-11-20). Per-supplier quotes and upstream self-check records are on the marketplace page.
OpenAI protocol: set base_url to https://321ai.cc/v1 and use the key generated in your console; model name is any of the Gemini models listed above. Existing SDKs need no code changes.
Multiple suppliers quote every model, so smart routing avoids failing channels automatically; merchants publish upstream self-check samples, and failed requests are never billed.
Registration includes free trial credit — call real models before you top up. Sign up: https://321ai.cc/ui/new/#/register