Gemini 3.5 Thinking API
Google reasoning for coding agents with metered billing.
Run gemini-3.5-thinking through a drop-in OpenAI-compatible endpoint. Coding Plan price is $0.90 input / $5.40 output per million tokens, billed through one Hypereal API key.
Independent third-party API aggregator. Not affiliated with or endorsed by any model provider; model names are trademarks of their respective owners.
Gemini 3.5 Thinking pricing
Hypereal public token pricing
Integrate in minutes
Standard REST API that works with any language. One API key gives you access to all models.
- Single endpoint for all models
- Bearer token authentication
- JSON request & response
- Webhook callbacks for async jobs
- Python & Node.js SDK available
from openai import OpenAI
client = OpenAI(
api_key="hyp-...",
base_url="https://api.hypereal.cloud/v1",
)
resp = client.chat.completions.create(
model="gemini-3.5-thinking",
messages=[{"role": "user", "content": "Review this diff and find the highest-risk bug."}],
)
print(resp.choices[0].message.content)Why Gemini 3.5 Thinking
Coding Plan eligible
Spend Coding Plan credits on Gemini 3.5 Thinking alongside Claude Opus, Sonnet, and GPT-5.5 — one prepaid pool, one API key.
Current metered pricing
$0.90 input / $5.40 output per million tokens through Hypereal Credits. No Google Cloud project or separate billing account required.
OpenAI-compatible
Use the same SDK shape as /v1/chat/completions; just set model to gemini-3.5-thinking.
Frequently asked questions
Is Gemini 3.5 Thinking included in the Coding Plan?
Yes. It is eligible for Coding Plan credits, so prepaid coding credits spend on it just like Claude Opus, Sonnet, and GPT-5.5.
What is the model ID?
Use gemini-3.5-thinking in chat, /v1/chat/completions, and any OpenAI-compatible SDK call.
How do I try it for free?
Every new account gets a free signup credit scoped to coding models — enough to make your first Gemini 3.5 Thinking call from Claude Code or Codex without paying.
Use Gemini 3.5 Thinking with Coding Credits.
Create an API key, set base_url to Hypereal, and call gemini-3.5-thinking from any OpenAI-compatible SDK.