Microsoft's Decision-1 model enters the fast-growing AI decision model race
Microsoft's Decision-1 hits 83.5% accuracy at 85ms latency—a specialized model for fast classification entering a crowded routing-and-classification space.
Why it matters
Microsoft ships a specialized small model (9B, built on Qwen3.5) optimized for low-latency classification and routing tasks. This reflects a shift toward task-specific model architectures rather than general-purpose agents, with real latency/accuracy tradeoffs practitioners need to evaluate against larger models and existing routing solutions.
The key facts
10 to knowModel: Decision-1, built on Qwen3.5-9B
Performance: 83.5% accuracy across 36 benchmarks (Microsoft's own tests)
Latency: 85ms
Use case: fast classification and routing
Status: appears to be a product release or announcement (GA implied, not explicitly stated)
Decision-1 built on Qwen3.5-9B base model
83.5% accuracy reported across 36 benchmarks
85ms latency for classification and routing
Optimized for fast classification and routing workflows
Company's own tests (not independently verified)
Go to the source
The Decoderthe-decoder.com
Publisher excerpt: With Decision-1, Microsoft enters the growing decision model space. Built on Qwen3.5-9B and optimized for fast classification and routing, it hits 83.5 percent accuracy with 85 ms latency across 36 benchmarks, according to the company's own tests.