Gemini 4 Carbon API is half price at $2 per million input tokens

Gemini 4 Carbon is Google's most capable reasoning model, built for complex software engineering, long-running agent work, and cybersecurity defense, with a 1M-token context window. It's now open on the API and billed at half price through November 30, so this is a good window to try it on a large codebase or a long agent workflow.

What you get

• Through November 30, the paid Gemini 4 Carbon API costs $2 per million input tokens and $10 per million output tokens, including thinking tokens. From December 1 it goes back to $4 input and $20 output.
• Context caching is half price too, at $0.10 per million tokens through November 30 and $0.20 after that.
• The price applies on the paid tier of Google AI Studio and the Gemini API. Vercel AI Gateway also carries the model at 50% off, and AI Gateway passes provider pricing through with no markup and no platform fee.
• Gemini 4 Carbon isn't available on the Free tier, so you need paid usage on the Gemini API or a funded Vercel AI Gateway account.
• There's no coupon or code; the discounted rate is applied automatically.

How to use

1. Turn on paid usage in Google AI Studio or the Gemini API and pick the model gemini-4-carbon.
2. Or sign in to Vercel, create an AI Gateway API key, and use the model name google/gemini-4-carbon. Running npx vercel@latest ai-gateway setup also wires it into Claude Code, Codex, Cursor, and other coding agents.
3. To try it first, open the Playground on the Vercel model page.

See Gemini API pricing (Gemini 4 Carbon), Google AI Studio, the Vercel AI Gateway model page, and Vercel's launch post.