Google · llm model

Gemini Decision 1

Gemini Decision 1 is Google's latest language model in our dataset, released Nov 11, 2026. Proprietary, API access only. Google's multimodal decision model behind the new Gemini Decisions API: it scores yes/no, multiple-choice and rubric questions in one pass instead of generating text. Announced Nov 11, 2026 and available in the Gemini API and Google AI Studio. Verified against vendor sources on Nov 12, 2026.

Reported Last verified Nov 12, 2026 · All models · JSON · Markdown

Primary source: https://blog.google/innovation-and-ai/technology/developers-tools/gemini-decisions-api/. We are the fast, independent index; the vendor’s own page is the source of record for these model facts.

Gemini Decision 1 at a glance: API id gemini-decision-1; 262,144-token context window; takes text, image, video, audio, pdf in; returns calibrated probabilities (yes/no, choice, score); listed price .

ReleasedNov 11, 2026
Context window262K262,144 tokens
LicenceProprietary

InTextImageVideoAudioPDFOutcalibrated probabilities (yes/no, choice, score)

Key facts

Vendor-reported facts are read from the linked vendor page on the date shown above; we did not run the model. Independently tested means we ran it; no fact on this page is. Editorial is our own classification, not a vendor claim. Where a linked page does not itself state a value, the row says so. How to read provenance.

Gemini Decision 1: key facts with sources
Fact	Value	Label	Source
Vendor	Google	Editorial	n/a
Released	Nov 11, 2026 Announced and available in the Gemini API (Google AI Studio) on Nov 11, 2026.	Vendor-reported	source
Status	generally available	Editorial	n/a
API id	gemini-decision-1	Vendor-reported	source
Context window	262,144 tokens (262K) Input token limit; output token limit is 0 (answers are probabilities, up to 1,000 questions per request).	Vendor-reported	source
Input	text, image, video, audio, pdf	Vendor-reported	source
Output	calibrated probabilities (yes/no, choice, score)	Vendor-reported	source
Licence	Proprietary	Vendor-reported	source
Pricing	Paid tier $0.10 per 1M input tokens (text, image, video, audio, PDF); output is not applicable because the model returns 0 output tokens. Free tier available. Shared input is billed once per request; each question adds its own instruction tokens. No context caching or Batch API. Prices: ai.google.dev/gemini-api/docs/pricing · Checked Nov 12, 2026	Vendor-reported	source

Where Gemini Decision 1 sits in our data

Released Nov 11, 2026 (week 46 of 2026), in November 2026. Other models in our data released that month: Gemini Robotics 2.5, Live 3, PAN 2, Scribe v3.
The previous Google model we track is Gemini Robotics 2.5 (Nov 10, 2026), 1 day earlier.
Of the 74 models in our data with a published context window, Gemini Decision 1 has the number 49 largest (262K tokens). Our data is not every model on the market.
Google tools we cover: Gemini Notebook, Google ADK, Gemini CLI.
All releases by date: AI model releases by month · new AI models this week. Choosing for code: best LLM for coding. Price and rank context: LLM pricing comparison · cheapest LLM API · OpenRouter pricing · best AI model right now.

Alternatives to Gemini Decision 1

Language models from other vendors that share at least one of our groups with Gemini Decision 1. Comparable on paper, from each model’s sourced facts; we have not run them.

Muse EmberMeta · Both in: Small and fast, Agents · $0.1 in / $0.5 out per 1M tokens · 1.05M context · released Nov 9, 2026
GPT-6.1 TerraOpenAI · Both in: Agents, Small and fast · $1 in / $5 out per 1M tokens · 1.05M context · released Oct 26, 2026
Mercury 3Inception · Both in: Small and fast, Agents · $0.1 in / $0.35 out per 1M tokens (launch promo, end date not stated; list $0.2 / $0.7) · 1M context · released Oct 19, 2026
MiniMax M3.1 FlashMiniMax · Both in: Agents, Small and fast · $0.15 in / $0.6 out per 1M tokens · 1M context · released Oct 19, 2026
Claude Haiku 5.5Anthropic · Both in: Small and fast, Agents · $0.1 in / $0.5 out per 1M tokens · 1M context · released Oct 7, 2026
Claude Fable 5.2Anthropic · Both in: Agents · $8 in / $40 out per 1M tokens · 1M context · released Nov 9, 2026

Side by side: Best AI model for agents (Nov 2026): Claude Opus 5.5 leads Terminal-Bench 4.0

How Gemini Decision 1 differs from Gemini 3.1 Flash-Lite

Read from the recorded facts of the two models, not from a head-to-head test. We have not run either.

We have a recorded price for Gemini Decision 1 but none for Gemini 3.1 Flash-Lite.
Context window is 262K tokens against 1.05M for Gemini 3.1 Flash-Lite.
Both accept text, image, video, audio, pdf as input.
Output is calibrated probabilities (yes/no, choice, score), against text.
Gemini Decision 1 came out 188 days after Gemini 3.1 Flash-Lite.
Facts for the other model: Gemini 3.1 Flash-Lite.

What the sources say

These are claims from the linked pages, not our measurements.

One request takes shared input up to 262,144 tokens (text, images, video, audio, PDF) and up to 1,000 independent questions; it returns probabilities and a confidence score, with 0 output tokens.
Vendor claim (Google): 150-250 ms end-to-end latency (50-100 ms of model compute) for inputs up to 16K tokens, and nearly 5x cheaper on Google's benchmark suite than Gemini 3.5 Flash-Lite with the Interactions API. We did not verify this.

Notes

Not a text generator: Google's model page marks text generation, function calling, structured outputs, thinking, caching and the Batch API as not supported. It is only used through the Decisions API.

What we tested

Nothing yet. We have not run Gemini Decision 1 ourselves, so every fact on this page is Reported from the sources below. When we do test it, the scope will be stated here and in our test logs. See methodology.

Not verified, left out

We could not confirm these on a vendor page, so we do not state them.

Benchmark scores: Google says Gemini Decision 1 leads its own suite with the lowest calibration error but does not name the compared models or publish the scores in the announcement.
Rate limits and free-tier quotas: the model page links to the rate-limits page; we did not record numbers.
Stability: Google's sample code uses the v1beta API version and no separate general-availability or deprecation date is stated.