---
type: "post"
id: "liquid-ai-released-d1-ultra-a-multimodal-decision-3c7286"
slug: "liquid-ai-released-d1-ultra-a-multimodal-decision-3c7286"
title: "Liquid AI released d1-ultra , a multimodal decision model that returns calibrated probabilities instead of generated text."
description: "Thesis: d1-ultra makes a probability-bearing decision interface a product surface alongside generated-text models."
retrieval_nugget: "Thesis: d1-ultra makes a probability-bearing decision interface a product surface alongside generated-text models."
published_at: "2026-10-16"
updated_at: "2026-10-16"
record_date: "2026-10-16"
date_kind: "published_at"
topics: ["multimodal"]
source_urls: ["https://www.liquid.ai/blog/d1-ultra"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:liquid-ai-released-d1-ultra-a-multimodal-decision-3c7286"
status: "published"
visuals: []
editorial_provenance: {"schema_version":"newruntime-editorial-copy-v1","content_status":"source_grounded_final","final_copy_sha256":"sha256:273c4826ac9a7bffdef4e1bfd56e37a7fc39bba8bf91fbdae9545f7df2a85de6","reviewed_at":"2026-10-16T08:01:32.976Z","source_evidence_count":1,"verified_claim_count":4,"site_analysis_schema_version":"newruntime-site-analysis-v1","site_object_kind":"field_note","observed_fact_count":4,"implication_count":1,"watch_condition_count":1,"related_record_count":0}
analysis: {"schema_version":"newruntime-site-analysis-v1","object_kind":"field_note","thesis":"d1-ultra makes a probability-bearing decision interface a product surface alongside generated-text models.","observed_facts":[{"text":"Liquid AI says d1-ultra accepts text, images, or both and returns probabilities without generating tokens.","source_urls":["https://www.liquid.ai/blog/d1-ultra"]},{"text":"The model supports yes/no, choice, and score question types.","source_urls":["https://www.liquid.ai/blog/d1-ultra"]},{"text":"Liquid AI lists API availability at $0.042 per million input tokens, with no output-token charge.","source_urls":["https://www.liquid.ai/blog/d1-ultra"]},{"text":"Liquid AI says it compared d1-ultra with d1 and Jev 1.14 across 112 evaluations.","source_urls":["https://www.liquid.ai/blog/d1-ultra"]}],"mechanism":"Rather than asking a language model to formulate a prose answer and then extracting a decision from it, the API takes one or more questions and returns probability values for bounded answer forms.","why_now":"The release gives builders an immediately accessible API and a stated price, so this decision-oriented interface can be evaluated in concrete workflows rather than treated only as a research direction.","implications":["Teams with bounded classification or scoring tasks can test whether direct probability outputs reduce the need to generate and parse prose."],"evidence_boundary":"Those benchmark and calibration statements come from Liquid AI’s own announcement; the supplied source does not independently validate them or establish performance in production.","watch_conditions":["Watch for independent evaluations of calibration and accuracy on representative real-world tasks, alongside confirmation of how API costs behave for multi-question and image-heavy inputs."],"related_records":[],"new_branch_reason":"No prior New Runtime record is supplied for this specific probability-output model release, so it starts a distinct analytical record about decision interfaces."}
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/liquid-ai-released-d1-ultra-a-multimodal-decision-3c7286/","markdown":"https://newruntime.com/posts/liquid-ai-released-d1-ultra-a-multimodal-decision-3c7286.md","json":"https://newruntime.com/posts/liquid-ai-released-d1-ultra-a-multimodal-decision-3c7286.json"}
related_patterns: ["model-answers-become-task-interfaces"]
related_shifts: ["generated-interface-from-answer"]
---

# Liquid AI released d1-ultra , a multimodal decision model that returns calibrated probabilities instead of generated text.

## Retrieval answer

Thesis: d1-ultra makes a probability-bearing decision interface a product surface alongside generated-text models.

Thesis: d1-ultra makes a probability-bearing decision interface a product surface alongside generated-text models. Liquid AI says d1-ultra accepts text, images, or both and returns probabilities without generating tokens. It supports yes/no, choice, and score question types. The model is available through the Liquid AI API at $0.042 per million input tokens, with no output-token charge.

Mechanism: Rather than asking a language model to formulate a prose answer and then extracting a decision from it, the API takes one or more questions and returns probability values for bounded answer forms. The company says images and text are billed at the same input-token rate and that each question is billed as its own prompt. Why now: The release gives builders an immediately accessible API and a stated price, so this decision-oriented interface can be evaluated in concrete workflows rather than treated only as a research direction.

Liquid AI reports comparisons with d1 and Jev 1.14 across 112 evaluations and claims better-calibrated answers and higher accuracy than Jev 1.14 at the same price. Evidence boundary: Those benchmark and calibration statements come from Liquid AI’s own announcement; the supplied source does not independently validate them or establish performance in production. First implication: Teams with bounded classification or scoring tasks can test whether direct probability outputs reduce the need to generate and parse prose. First watch condition: Watch for independent evaluations of calibration and accuracy on representative real-world tasks, alongside confirmation of how API costs behave for multi-question and image-heavy inputs.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Multimodal](https://newruntime.com/topics/multimodal.md): Explore the multimodal topic hub.
- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)