Liquid AI d1-ultra: SOTA decision model with zero output tokens
Liquid AI released d1-ultra, a multimodal decision model that takes text/images plus typed questions and returns calibrated probabilities in one forward pass with zero output tokens, at $0.042 per million input tokens.
Why it matters
Ranks #1 on the independent Artificial Analysis Decision Index at 65 with the lowest calibration error Liquid tested, undercutting general-LLM classification on cost and latency.
What you could build with it
A trust-and-safety team could sit d1-ultra in front of a live support queue to triage tickets into risk buckets, routing only high-uncertainty cases to human reviewers and cutting review volume while keeping the audit trail of calibrated scores.
Does it hold up?
Available today via the Liquid API with a playground; benchmark claims are vendor-reported on vendor-chosen tasks, and no independent deployments have been published yet.
Built with Liquid AI d1-ultra
- MarkTechPost: Liquid AI releases d1-ultrablog · Technical coverage of the announcement: API structure, zero output tokens, and the independent Artificial Analysis Decision Index score of 65.
- ExplainX: Liquid AI d1-Ultra — a $0.042 decision model claiming the top spotblog · Independent writeup separating Liquid's claims from what can be checked on the benchmark scores.
Learn more
First spotted on article: source.
More AI models
OpenAI publishes 722 AI-generated math manuscripts from an unreleased modelOpenAI released 722 mathematical manuscripts in 372 result families to GitHub under Apache-2.0, produced by an…models · JEV 0.74StepFun Step-5-Preview — open BF16 weights releasedStepFun published the full BF16 weights for its flagship Step-5-Preview agentic model (600B-parameter sparse MoE, 27B…models · JEV 0.73ElevenLabs Eleven v4 + v4 Turbo: new emotive TTS architecture with a 100ms real-time variantElevenLabs released Eleven v4 and Eleven v4 Turbo on September 28: a new text-to-speech architecture with inline…models · JEV 0.72Mistral Large 4 (Le Chonk): 1.05T-param open-weight MoEMistral's new flagship: a 1.05T-parameter mixture-of-experts multimodal model with 49B active params and 1M-token…models · JEV 0.72VoiceStudio: open-source, fully-local ElevenLabs alternative with voice cloning, dubbing and dictation in 646 languagesVoiceStudio is an open-source, fully-local voice AI studio: voice cloning from a 3-second sample, voice design, video…models · JEV 0.71JetBrains Mellum 2.1: 12B MoE coding model trained for agentsJetBrains released Mellum 2.1, a 12B MoE model (2.5B active parameters, 128K context) under Apache 2.0, trained mainly…models · JEV 0.69
Get the week's best AI launches, plus 3 ideas worth building
One email every Saturday. Ranked by traction, not hype. Free.