Liquid AI's d1 is available on AI Gateway
Liquid AI's d1 decision model is now accessible through Vercel's AI Gateway—structured classification without text generation, with vision support.
Why it matters
Liquid d1 adds a decision-model alternative to Vercel's unified AI Gateway, enabling typed classification, routing, and scoring without token generation. Practitioners can route support tickets, validate refunds, or perform visual inspection through SDK, OpenAI-compatible, or TypeSafe APIs. Decision calls count toward budgets and appear in logs alongside other model requests.
The key facts
13 to knowLiquid AI d1 now available on Vercel AI Gateway
d1 evaluates shared state against typed questions for classification, routing, and scoring
Returns structured answers with probabilities, no text token generation
Vision support: visual classification, inspection, and scoring
Accessible via SDK, OpenAI-compatible Decisions API, or TypeSafe-compatible API
Decision calls tracked in logs and budgets alongside other model requests
Examples: support-agent refund validation, local image color classification
Liquid d1 available on Vercel AI Gateway
Model type: decision model for classification, routing, scoring without text generation
Vision support: can answer typed questions about images
Access methods: AI SDK, OpenAI-compatible Decisions API, TypeSafe-compatible API
Use cases: support agent refund classification, visual inspection, color detection
Published: October 9, 2026
The story so far
Earlier coverage of this storyline
- FLUX 3 Image now available on AI GatewayVercel Blog
- Mistral Large 4 now available on AI GatewayVercel Blog
- OpenAI Decisions API now available on AI GatewayVercel Blog
- Claude Haiku 5.5 now available on AI GatewayVercel Blog
- Step 5 Preview now available on AI GatewayVercel Blog
- This story
Go to the source
Vercel Blogvercel.com
Publisher excerpt: is now available on .Liquid d1AI Gateway d1 is a that evaluates shared state against typed questions for classification, routing, and scoring. It returns structured answers with probabilities without generating text tokens. d1 also has vision support: the model can answer typed questions about…