Skip to content

ModelsLiquid AI

Liquid AI releases d1.2 and d1.2-ultra: decision models that add audio input and five new typed tasks

Liquid AI d1.2 at a glance

As of November 16, 2026, d1.2-ultra ranks #1 of 24 decision models on the Artificial Analysis Decision Index v2.1, at 74.8. Liquid AI's new d1.2 decision models take text, images and audio, return calibrated probabilities with zero output tokens, and bill input only, from $0.02 per 1M tokens.

Ask your AI about this story

Use this story in your business: Liquid AI releases d1.2 and d1.2-ultra: decision models that add audio input and five new typed tasks (Liquid AI). Ask your AI to explain what happened, how it could affect your business and how to put it to work. It also weighs price, speed, intelligence and the lab behind it. The link to the story is included.

What changed with Liquid AI d1.2

Liquid AI released d1.2 and d1.2-ultra on November 16, 2026, the new versions of its hosted decision models, on the Liquid API. A decision model does not generate text: you give it a state — a document, a picture, a recording — plus typed questions, and in a single pass it returns calibrated probabilities or structured answers, with zero output tokens. Both models now take audio as input, and the question itself can be spoken instead of typed. They add five typed tasks: read a value from a document and cite the characters it came from, or answer not found; find every passage of a kind, such as personal data, with its position; locate objects in a picture with boxes and a count; estimate a number with a range; and pick every option that applies, each with its own probability.

As of November 16, 2026, Artificial Analysis scores d1.2-ultra at 74.8 on its Decision Index v2.1, #1 of the 24 dedicated decision models it tracks, at $133 per 1M tasks and a 218 ms median time per task; d1.2 scores 68.7. Artificial Analysis's v2 article measured general language models higher on the same index — Claude Haiku 5.5 (Max) at 77 and GPT-6 Luna (Max) at 76 — at far higher cost per task. Pricing is input tokens only: $0.02 per 1M for d1.2 and $0.042 per 1M for d1.2-ultra, with images billed at 1.5 tokens per 32×32 patch. Liquid AI says the models are coming to OpenRouter and Vercel soon, and reports a 0.8-point calibration error for d1.2-ultra across 13 benchmarks in its own tests.

Liquid AI d1.2 benchmarks and pricing

Values are copied from the linked source on the date shown. They are not IA Compa measurements.

Artificial Analysis Decision Index v2.1 score · d1.2-ultraas of · Source for Artificial Analysis Decision Index v2.1 score: artificialanalysis.ai (opens in a new tab)
74.8
Artificial Analysis Decision Index v2.1 scoreas of · Source for Artificial Analysis Decision Index v2.1 score: artificialanalysis.ai (opens in a new tab)
68.7
Input price per 1M tokens · d1.2-ultraas of · Source for Input price per 1M tokens: docs.liquid.ai (opens in a new tab)
$0.042
Input price per 1M tokensas of · Source for Input price per 1M tokens: docs.liquid.ai (opens in a new tab)
$0.02
Cost per 1M tasks on the Decision Index (Artificial Analysis) · d1.2-ultraas of · Source for Cost per 1M tasks on the Decision Index (Artificial Analysis): artificialanalysis.ai (opens in a new tab)
$133
Median time per task on the Decision Index (Artificial Analysis) · d1.2-ultraas of · Source for Median time per task on the Decision Index (Artificial Analysis): artificialanalysis.ai (opens in a new tab)
218 ms

How it compares

Current models we have covered: released in the last six months, newest version of each line only. Each value is the model's latest recorded value. (Models)

Input price (USD per 1M tokens) · d1.2-ultralower is better
Input price (USD per 1M tokens) · d1.2lower is better

Where to compare Liquid AI d1.2

Independent leaderboards and official resources for comparing this model yourself. Rankings change often; each site shows its own current data.

Why Liquid AI d1.2 matters for businesses

Why it matters: the classify, route and extract steps inside business automations — sorting support tickets, triaging leads, pulling totals off invoices, flagging personal data before a document is stored — can now run on a purpose-built decision model at $0.02–$0.042 per million input tokens and millisecond latency, instead of a full generative model call. Because every answer carries a calibrated probability, a business can auto-approve the confident cases and route the rest to a person, and the new span, box and audio tasks extend that pattern to scanned documents, photos and call recordings.

Put it to work: AI services

Get AI News by email

Verified AI stories, each with its official source, on Monday and Thursday mornings. Add breaking news to hear about the biggest ones as soon as we publish them.

Every story links its official source

Free, in English or Spanish

Unsubscribe with one click, any time

What to send

Sources for this story

  1. Official
  2. Official
  3. Artificial Analysis Decision Index (opens in a new tab)

    Artificial Analysis · artificialanalysis.ai

    Benchmark
  4. Benchmark

Last verified: . IA Compa summarizes public announcements for business readers. We are not affiliated with the companies mentioned. Always confirm details, pricing, and availability with the official source.

  • #liquid-ai
  • #d1-2
  • #d1-2-ultra
  • #decision-models
  • #artificial-analysis-decision-index
  • #calibrated-probabilities
  • #classification
  • #extraction
  • #audio-input

Want this working in your business?

We turn new AI capabilities into agents, voice receptionists, and automations you can actually use.