Google’s Gemini Decision 1 does not write answers. It returns probabilities: you give it a question and the options, and it tells you which option fits and how sure it is, according to Google’s model documentation.
Google lists its input price at $0.10 per million tokens, with zero output tokens, on its pricing page. Gemini 4 Hydrogen’s input costs $1.50 per million, 15 times more, by our arithmetic. It is a new line on our AI model benchmarks page.
Google released the model and its Decisions API on 11 November, its release notes show. Google for Developers India promoted it on X on Saturday.
Business Pill · The Model and the Harness
An evergreen Business Engineer explainer. It teaches the general idea only and does not discuss this model.
The key insight: Gemini Decision 1 returns probabilities instead of text, at $0.10 per million input tokens and zero output tokens, Google says; Gemini 4 Hydrogen’s input price is 15 times that, by our arithmetic.
What It Does
Most AI models generate text word by word, and your code then has to read that text to decide what to do. Gemini Decision 1 skips the writing. It evaluates typed questions in a single pass and returns calibrated probabilities, Google says.
It handles three kinds of question, according to Google’s documentation. A yes/no question returns a probability from 0.01 to 0.99 plus a confidence score. A choice question picks the top option from 2 to 255 options. A rubric question returns an expected score across 2 to 255 ordered levels.
It reads text, images, video, audio and PDFs. One request can ask up to 1,000 separate questions about the same input, which can run to 262,144 tokens, Google says.

What It Costs
Google bills Gemini Decision 1 at $0.10 per million input tokens on its paid tier, and there are no output tokens to pay for. The shared input is billed once however many questions you ask, and each question adds its own instruction tokens, the pricing page says.
Gemini 4 Hydrogen, Google’s everyday model, lists $1.50 per million input tokens and $7.50 per million output tokens. Gemini 4 Neon lists $3.00 and $15.00. Hydrogen’s input price is 15 times Decision 1’s and Neon’s is 30 times, by our arithmetic.
A worked example, by our arithmetic on list prices: sorting one 1,000-token customer message would cost $0.0001 with Decision 1. With Hydrogen writing a 20-token answer, it would cost about $0.00165, roughly 16.5 times as much, before any thinking tokens.
What It Is For
Google names request routing, policy checks across many rules at once, scoring candidates against a rubric, tool selection and fast classification over live media streams as uses. Its example sorts a complaint about a cracked mug into a replacements queue.
The confidence score is meant to tell your code whether to act automatically or pass the case to a person, Google’s Decisions API guide says. Each question runs on its own against the shared input, so questions cannot see each other.
What It Does Not Do
It does not generate text, call functions, search the web or think step by step; Google’s model page lists all of those as not supported. Batch processing and caching are not supported either. It is available through the Gemini Developer API in a v1beta version.
The Structural Read
As we read it, this is a model built for software to call, not for people to read: agents and apps that need a yes, a no or a choice can skip writing and parsing text.
As we read it, pricing judgment calls this low pushes more routine routing and moderation into AI rather than hand-written rules.
Google AI Studio docs
“It generates zero text tokens, so output token usage is always 0.”
Three Implications
FOR DEVELOPERS one request can ask up to 1,000 questions about the same input, Google says.
FOR COST input is $0.10 per million tokens with no output tokens, per Google’s pricing page.
FOR LIMITS no text generation, function calling, search or thinking, per Google’s model page.
What Is Not Established
Google’s pages describe the probabilities as calibrated but, in what we read, give no accuracy benchmark or latency figure. We have not seen independent tests of the model. The prices are list prices and can change.
The Bottom Line
Gemini Decision 1 answers questions with probabilities instead of words, at $0.10 per million input tokens and no output tokens, Google says. That input price is one-fifteenth of Gemini 4 Hydrogen’s, by our arithmetic.
97,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. Every figure comes from Google’s own AI Studio documentation, pricing page and release notes, read on 21 November 2026. The 15 times, 30 times and the worked example are our arithmetic on list prices; they ignore free tiers, discounts and thinking tokens. Google has published no independent benchmark of the model that we have seen.
Sources: Google AI Studio documentation, ‘Gemini Decision 1’ model page, read 21 Nov 2026 · Google AI Studio documentation, ‘Decisions API’ guide, read 21 Nov 2026 · Google AI Studio pricing page (Gemini Decision 1, Gemini 4 Hydrogen, Gemini 4 Neon), read 21 Nov 2026 · Google AI Studio release notes, entry dated 11 Nov 2026 · Google for Developers India on X, 21 Nov 2026









