AINovaTools

Last updated: 10/11/2026Last verified: 2026-10-10

Microsoft-Decision-1

Microsoft-Decision-1

4
0

Microsoft's small decision-scoring model available in Microsoft Foundry and via OpenRouter: given text or JSON state and typed questions (yes/no, choice, score), it returns calibrated probabilities instead of generated text, for classification, routing, prioritization, verification, and workflow control.

Paid
AD

<p>Microsoft''s small decision-scoring model available in Microsoft Foundry and via OpenRouter: given text or JSON state and typed questions (yes/no, choice, score), it returns calibrated probabilities instead of generated text, for classification, routing, prioritization, verification, and workflow control.</p><ul><li>Returns calibrated probability per fixed answer option instead of free-form generated text</li><li>Typed noul (yes/no), choice, and score question types for gates, filters, routing, and grading</li><li>Available in Microsoft Foundry with Entra ID or API-key auth, and through the OpenRouter Decisions API</li><li>Post-trained from Alibaba open-weight Qwen3.5-9B for single-pass scoring; weights update while the API shape stays stable</li></ul>

Details

Platform

API

Features

  • Returns calibrated probability per fixed answer option instead of free-form generated text
  • Typed noul (yes/no), choice, and score question types for gates, filters, routing, and grading
  • Available in Microsoft Foundry with Entra ID or API-key auth, and through the OpenRouter Decisions API
  • Post-trained from Alibaba open-weight Qwen3.5-9B for single-pass scoring; weights update while the API shape stays stable

Languages

en

Best for

Applications that need predictable, typed decision outputs with confidence thresholdsSupport-ticket, incident, and content triage flows: classify, route, prioritize, and filterAgent guardrails and AI judging: verification, prompt-injection checks, and model/agent selection

Known limitations

  • Not intended for open-ended generation, conversation, translation, or summarization
  • Uses the OpenRouter Decisions API rather than the OpenAI-compatible chat endpoint, so standard chat SDKs do not work
  • OpenRouter pricing is $0.042 per 1M input tokens with output free; throughput is capped by the single Azure provider
  • Probability thresholds for automated gates should be validated against labeled examples before use

Alternatives

Jev — TypeSafe decision API that also returns typed probability-calibrated answers
OpenAI Structured Outputs — General LLM API for structured JSON when decisions also need reasoning text
Anthropic Claude API — General model API for richer classification workflows with tool use
View all Microsoft-Decision-1 alternatives →

Rate This Tool

Related Tools

ScaleDown
ScaleDown
Paid
Task-specific AI API for classifying content, extracting structured fields, summarizing documents, and compressing agent context. Text input costs $0.05 per million tokens, supported image input costs $0.15 per million, and output is free; the current launch offer gives new accounts 50 million free tokens.
ai-codingai-productivity
1
0
Prime Agent
Prime Agent
Free
Prime Intellect's open-source, Rust-rewritten coding and research agent harness that uses recursive subagents, a persistent Python REPL, daemon-backed sessions, and a Continual Harness — with 14x faster input latency, 80%+ lower memory, and new Windows beta support.
ai-codingai-productivity
6
0
Mem0
Mem0
Freemium
Mem0 is an open-source AI memory layer that lets assistants and agents remember user preferences across sessions, with a pip/npm library, a self-hostable server, and a managed cloud platform.
ai-productivityai-coding
3
0
Seesa
Seesa
Subscription
Seesa is on-premise company AI from UK maker Pentatonic: Qwen-family models, a company memory over email and documents, and approval-gated agent work installed in your building and managed for you, with SOC 2 Type II coverage.
ai-productivityai-search
4
0
Agent Skills
Agent Skills
Free
Agent Skills is an MIT-licensed collection of production-grade workflows for AI coding agents, with lifecycle commands and reusable skills covering specification, planning, implementation, testing, review, and shipping.
ai-codingai-productivity
2
0
Prime Inference
Prime Inference
Paid
Production inference platform from Prime Intellect offering OpenAI-compatible access to hosted and gateway models, with serverless endpoints, reserved capacity, multi-datacenter failover, and infrastructure for long-running agent workloads.
ai-codingai-productivity
2
0