Tasks and inputs
Microsoft-Decision-1.1 is listed for Image classification, Text classification, Structured Decisions. It accepts text, image and produces text. Use the publisher’s input format and preprocessing for your chosen deployment.
SyncDev model directory · Profile updated Nov 18, 2026
Microsoft-Decision-1.1 is a decision model for routing, triage and classification, returning calibrated probabilities. GA in Foundry from November 2026.
The specification score summarizes recorded features and limits. It is not a benchmark result.
Pricing conditions: Indicative: from $0.042 per 1M input tokens (Global Standard) in launch-week listings; confirm in the Foundry catalogue for your region.
Recorded modality support shows what this model can accept and produce. An undocumented modality is shown as unknown rather than assumed to be unsupported.
Accepts
Produces
Not every step in an AI pipeline needs a model that writes. Many need one that chooses: which agent should take this request, is this ticket urgent, does this screenshot show an error. Microsoft-Decision-1.1, made generally available on 17 November 2026, is built for that narrower job. It picks among answers you define and returns calibrated confidence for each, rather than generating prose.
Microsoft uses it on its own traffic. Copilot relies on it to route requests between Copilot and Cowork, Cowork's Auto mode uses it to choose how much reasoning effort a task deserves, and Foundry Evaluations uses it as a judge.
That last point is the design. You decide what 0.7 means for your business.
Decision-1 entered public preview in October 2026 and picked up more than 4,000 customers in about five weeks. Version 1.1 adds image input, so photos and screenshots can be part of the question. Microsoft also reports better calibration and more than twice the speed at the same benchmark cost. On the Artificial Analysis Decision Index v2.1 it reports a score of 68, six points above Decision-1. These are publisher-reported numbers.
Under the hood it is built on Gemma 4 26B A4B, a mixture-of-experts model, served on a custom inference engine.
| Detail | Value |
|---|---|
| Status | Generally available, 17 Nov 2026 |
| Inputs | Text and images |
| Output | Probabilities over defined answers |
| Platforms | Microsoft Foundry, OpenRouter, Vercel AI Gateway |
| Indicative price | From $0.042 per 1M input tokens |
The price comes from launch-week listings rather than a published rate card, so check the Foundry catalogue for your region before you budget.
The model turns up as "Decision 1.1", "Microsoft-Decision-1.1" and occasionally "MAI Decision 1". They refer to this family. Search the catalogue for "Decision" rather than a specific spelling.
Take a week of requests your current router or LLM classifier handled. Label a few hundred by hand. Run them through Decision-1.1 and plot accuracy against confidence: a well-calibrated model should be right about 90% of the time when it says 0.9. If it is, you can set an escalation threshold with real confidence. Compare it with Jev and other entries under structured decisions, and keep a generative fallback such as GPT-6 Luna for the cases it abstains on.
Compare the model on the work you need to complete.
Microsoft-Decision-1.1 is listed for Image classification, Text classification, Structured Decisions. It accepts text, image and produces text. Use the publisher’s input format and preprocessing for your chosen deployment.
The listed route has starting rates of $0.042 input per million tokens. Measure a completed task using the prompts, documents and tools your application will use. Include reasoning and retries in the result.
Use the model identifier and account requirements for the selected API route. Confirm its request limits and output format with one complete task before connecting a larger workflow.
These records share a directory category. Check the source card, output type and deployment requirements before treating a model as a replacement.
Liquid AI
Liquid AI d1.2 is a hosted omnimodal decision model for text, image and audio questions with calibrated answers, at $0.02 per million input tokens.
Liquid AI
Liquid AI d1.2-ultra is the top hosted decision model in Liquid's line, scoring 75 on the AA Decision Index, with audio support at $0.042 per 1M input.
Gemini Decision 1 powers Google's Decisions API, returning calibrated yes/no, choice and rubric answers instead of text, at $0.10 per 1M input tokens.
subquadratic
SubQ Snap 1.2 is Subquadratic's long-context decision model on linear sparse attention, tested to 256K tokens, priced at $0.04 per 1M input tokens.
Perplexity
pplx-decider-flash-v1.3 is Perplexity's small, half-price decision model, distilled from the 27B decider, for high-volume classification and span extraction.
Perplexity
pplx-decider-v1.3 is Perplexity's 27B-class decision model: it returns probabilities for choice, yes/no, score and new span-locate questions, not text.