Liquid AI launches d1 decision model for structured tasks with zero output tokens
Liquid AI released d1, a decision model designed for structured choices like classification, ticket routing, and moderation. Unlike text-generating models, d1 returns calibrated probabilities across predefined outcomes in a single API call, with no generated tokens billed. It supports three primitives—Noul (yes/no with probability), Choice (top pick from a set), and Score (ordered rubric…
Key points
- d1 returns typed decisions with calibrated probabilities and zero output tokens, billed only for input
- Supports three primitives: Noul (yes/no), Choice (top pick), and Score (ordered ratings) in one call
- Deployable now via Liquid’s API as **d1:free**; no self-hosted weights or training options
The model is available today via Liquid’s hosted API under d1:free, with no self-hosting options. Liquid’s migration guide advises using d1 for tasks with fixed answer sets (e.g., spam detection, urgency triage) and reserving LLMs for generation or multi-step reasoning. A demo shows d1 powering a real-time road-decision game, outperforming a competitor model in consistency. The company highlights benefits like predictable latency, no schema errors, and reduced round trips compared to LLMs.
Liquid AI Releases d1: A Decision Model That Returns Calibrated Probabilities With Zero Output Tokens
MarkTechPost · 29 September 2026
Loading the full article…
This text was published by MarkTechPost and written by Asif Razzaq. It is reproduced here with attribution so you can read it in full; the rights remain with the publisher. Read it at the source ↗
The headline, key points and digest above were generated by Digest AI's editorial model from the linked sources. Automated summaries can contain errors: the sources are the record. Spotted a mistake? Tell us. Published by Martin K., who runs Digest AI and handles corrections.
More in Generative AI & Models
All →- OpenAI faces concerns over loop transformers; Microsoft confirms GPT-6.1 Sol uses them · 1 src
- Meta introduces Muse Wonder and Muse Wonder Full models · 8 src
- Anthropic releases Claude Haiku 5.5 at lower prices than GPT-6 Luna · 1 src
- Microsoft releases Microsoft-Decision-1, a Qwen3.5-9B-based scoring model · 2 src
- Mistral releases Mistral Large 4, nicknamed Le Chonk, a 1-trillion-parameter multimodal model · 1 src
Comments
via GitHub Discussions