Google launches Gemini Decision 1 and a Decisions API: up to 1,000 yes/no, choice and rubric questions per call at $0.10 per 1M input tokens, with zero output tokens

Google launched Gemini Decision 1 (gemini-decision-1) and a new Decisions API in the Gemini API on November 11. The model does not write text. It scores fixed answers to typed questions (yes/no, multiple choice or a rubric) in one pass and returns probabilities with a confidence score. It costs $0.10 per 1M input tokens, bills zero output tokens, and has a free tier in Google AI Studio. Google's blog post went up at 18:00 UTC, and the Gemini API release notes, models page and pricing page all list it.

THE PRICE AND LIMITS

• $0.10 per 1M input tokens for text, image, video, audio and PDF input. There is no output price, because the model returns no output tokens. The free tier is free of charge.
• The shared input is billed once per request, however many questions ask about it. Each question adds only its own instruction and criteria tokens.
• One request can ask up to 1,000 questions about shared input of up to 262,144 tokens. Choice and rubric questions take 2 to 255 options or levels.
• No context caching and no Batch API.
• Rate limits: 200 requests a minute and 10,000 a day on the free tier, 1,500 a minute on Tier 1, and 15,000 a minute on Tier 3.

GOOGLE'S OWN CLAIMS

• End-to-end latency of 150 to 250 ms, with median model compute of 50 to 100 ms for inputs up to 16K tokens, nearly flat from 1 to 100 questions per request.
• Google says Decision 1 leads its own benchmark suite of agent routing, reasoning, ranking and multimodal tasks, with the lowest calibration error. It also says Decision 1 is nearly 5x cheaper across that suite than Gemini 3.5 Flash-Lite on the Interactions API. These are Google's figures, not an independent test.
• Google says its own products already use it, including model routing in Gemini Notebook, where it says everyday queries get up to 40% lower latency.

WHERE IT FITS

• It follows OpenAI's Decisions API, which went into public beta on October 6, and Microsoft-Decision-1, announced October 9.
• ArtificialWatch's radar logged the gemini-decision-1 id on Google's model list at 16:45 UTC, 75 minutes before Google's blog post.
• ArtificialWatch's launch alerts cover models that generate text, so this launch did not page subscribers.

WHAT WE DO NOT KNOW

How Decision 1 scores on independent decision benchmarks such as JevBench; Google's comparisons are its own. Google says integrations are rolling out on OpenRouter, LangChain, Vercel, LiteLLM and others. At 18:52 UTC OpenRouter lists no decision id. Google says support in Gemini in Google Cloud is coming soon, with no date.