Felo API PlatformFelo API Platform
Back to models

d1.2 API

LiquidAI d1.2 is a multimodal decision model that returns probability-based structured answers without generating output tokens.

LiquidAI d1.2 is the latest version of Liquid AI's decision model and part of the Liquid Foundation Models series. It processes text input, image input, audio input, or combinations of these modalities, then returns probabilities for multiple structured questions in one forward pass. Supported formats include yes/no questions, multiple-choice questions, and scale-position questions.

The model is designed for structured decisions rather than free-form text generation. It suits support-ticket filtering, intent and topic classification, fraud detection, routing and triage, moderation, guardrails, inspection workflows, product-image matching, and game-state decisions. It can also evaluate customer messages in real time for refund intent, routing destination, and urgency. Because it does not generate output tokens, text decisions complete in200〜300 milliseconds. d1.2 is available through the Liquid AI API, Vercel, and OpenRouter, although Vercel and OpenRouter currently support text-only use.

Updated:

MODEL ID

d1.2
Context
65.536K
Max output
58.982K
Input
text / image
Output
decisions

Highlights

  • Multimodal structured decisions
  • Multiple questions per pass
  • No output-token generation
  • Probability-based answers
  • Open-weight model
  • Edge and cloud deployment

Providers

1 providers

Compare API prices across providers. Select a provider to view all charges and conditions.

Provider · Pricing details
Uptime
$0.02$0

Benchmarks

  • Artificial Analysis Decision Index v2.1

    Ranked #1.

    75
  • Liquid AI calibration test

    Points of error.

    0.8

Scores as published by the model provider.

Decisions Quick Start

Send state and narrow typed questions to /api/v1/decisions and read one answer per question: noul returns a probability from 0 to 1, choice the selected option with its distribution, and score a numeric score for ordered criteria. This endpoint is separate from the Chat Completions, Messages, and Responses surfaces.

curl https://openapi.felo.ai/api/v1/decisions \
  -H "Authorization: Bearer $FELO_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "d1.2",
    "state": "Help! My payouts have been failing for 3 days.",
    "questions": {
      "is_urgent": {
        "type": "noul",
        "instructions": "Does this message convey urgency?",
        "criteria": {"true": "Time-sensitive", "false": "No urgency"}
      },
      "department": {
        "type": "choice",
        "instructions": "Which team should handle this?",
        "criteria": {
          "billing": "Payments, invoicing, refunds",
          "technical": "Bugs, outages, integrations",
          "sales": "Pricing, upgrades, new accounts"
        }
      },
      "frustration": {
        "type": "score",
        "instructions": "How frustrated is the customer?",
        "criteria": ["Calm", "Frustrated", "Very angry"]
      }
    }
  }'

Frequently asked questions

What is d1.2?

LiquidAI d1.2 is the latest version of Liquid AI's decision model and part of the Liquid Foundation Models series. It processes text input, image input, audio input, or combinations of these modalities, then returns probabilities for multiple structured questions in one forward pass. Supported formats include yes/no questions, multiple-choice questions, and scale-position questions. The model is designed for structured decisions rather than free-form text generation. It suits support-ticket filtering, intent and topic classification, fraud detection, routing and triage, moderation, guardrails, inspection workflows, product-image matching, and game-state decisions. It can also evaluate customer messages in real time for refund intent, routing destination, and urgency. Because it does not generate output tokens, text decisions complete in200〜300 milliseconds. d1.2 is available through the Liquid AI API, Vercel, and OpenRouter, although Vercel and OpenRouter currently support text-only use.

How is usage billed?

Your bill depends on the provider you use, its prices and your usage. Longer inputs or other conditions may have different prices.

What inputs does LiquidAI d1.2 support?

It supports text input, image input, audio input, and combinations of these inputs. Vercel and OpenRouter currently support text-only use.

What kind of output does d1.2 return?

It returns probability-based structured decisions for multiple questions in one forward pass. It supports yes/no, multiple-choice, and scale-position formats, and does not generate output tokens.

How fast are text decisions?

Text decisions complete in200〜300 milliseconds.

What tasks is d1.2 suitable for?

It is suited to classification, routing, triage, fraud detection, moderation, inspection, product-image matching, game-state decisions, and real-time customer-message evaluation.

How does d1.2 differ from d1.2-ultra?

d1.2-ultra is positioned as the world's best decision model for difficult real-world cases, while d1.2 is the latest version of Liquid AI's decision model.

Related models