decision-1 — API, Pricing & Context Window | Vivgrid

decision-1 is Microsoft-Decision-1, Microsoft's model for decision and classification workloads, and it does not write open-ended text. You send it the current state plus the questions you need answered, and it returns predefined answers with a confidence signal for each — so your application can act on confident answers and route uncertain ones for further review.

Microsoft post-trained it from Qwen3.5-9B for fast, single-pass decision scoring and released it in public preview on Microsoft Foundry on October 9, 2026. It is meant for simple, repeatable, time-sensitive decisions — agent controls, model routing, intent analysis, data labeling, AI judging — that don't need the generative capabilities or extended reasoning of a larger language model. Microsoft reports its median (P50) latency as about 35× faster than GPT-6 Sol.

On Vivgrid, decision-1 is served on the /systemone API with the same state + questions request shape as jev-latest, with geo-distributed acceleration across AMER, EMEA, and APAC. Output tokens are free.

Specifications

Pricing

Pricing in USD per 1M tokens.

Question types

• Yes/no — whether a statement about the state is true
• Multiple choice — pick one option from a fixed set
• Rating — place the state on an ordered scale
• Rubric grading — judge an AI response or an agent action against your criteria

A single request may carry several questions; on Vivgrid the whole request must fit in 32K tokens.

API surfaces

The Vivgrid Console catalogs decision-1 on one surface, all behind the same API key. The endpoint shown is the one the Console previews for that surface.

Quick start

Get an API key from the Vivgrid Console, then call decision-1 directly.

curl https://api.vivgrid.com/v1/systemone \
  -H "Authorization: Bearer $VIVGRID_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "decision-1",
    "state": "The website has been down for 2 hours and customers cannot complete purchases.",
    "questions": {
      "routing": {
        "type": "choice",
        "instructions": "Which team should handle this?",
        "criteria": {
          "billing": "Payments, invoicing, refunds",
          "technical": "Bugs, outages, integrations",
          "sales": "Pricing, upgrades, new accounts"
        }
      },
      "severity": {
        "type": "score",
        "instructions": "How severe is this issue?",
        "criteria": ["Minor", "Moderate", "Critical"]
      }
    }
  }'

Ideal use cases

• Agent controls and model routing between steps of a pipeline
• Intent analysis and incident-response routing
• AI judging: grading model output or agent actions against a rubric
• High-volume data labeling, data validation, and content classification

Related models

• jev-latest — TypeSafe's System One decision model on the same /systemone API
• gemini-3.8-flash — a fast general-purpose model when you need generated text
• glm-5.3-flash — low-cost chat completions for lightweight classification