We use cookies. Cookies and browser storage keep you signed in and remember your preferences. With your permission we also measure how the site is used. You can any time. See our cookie policy.

New here? See how it works A 1-minute tour of the playground, models and your API key.

LiquidAI: d1-ultra vs ScaleDown: Classify-1

A side-by-side comparison of LiquidAI: d1-ultra ($0.15 / 1M tokens) and ScaleDown: Classify-1 ($0.14 / 1M tokens) using current USD prices and capabilities from the api14 catalog. Both run through the same OpenAI-compatible API.

Specifications

	LiquidAI: d1-ultra	ScaleDown: Classify-1
Provider	LiquidAI	ScaleDown
Input	$0.15 / 1M tokens	$0.14 / 1M tokens
Cache read	$0.15 / 1M tokens	-
Context window	32,768	1,000,000
Max output	29,491	900,000
Input	text, image	text, image, file
Output	decisions	decisions
Tool calling
Structured output
Reasoning control
Added	Oct 15, 2026	Oct 13, 2026

Key differences

Both have about the same input price.
ScaleDown: Classify-1 reads 30.5× more context (1,000,000 tokens).
A request with 1,000 input and 500 output tokens costs $0.000147 with LiquidAI: d1-ultra and $0.000140 with ScaleDown: Classify-1.

In depth

LiquidAI: d1-ultra is a decisions model from LiquidAI, released on 15 October 2026. ScaleDown: Classify-1 is also a decisions model, made by ScaleDown and released two days earlier on 13 October 2026. Both sit in the same category, meaning both are built to return a decision as their output rather than free form prose, and both are served through api14 under the model ids liquid/d1-ultra and scaledown/classify-1.

The clearest split between them is context and input handling. ScaleDown: Classify-1 accepts a context window of 1,000,000 tokens, while LiquidAI: d1-ultra accepts 32,768 tokens, so the ScaleDown model is the one with the bigger context by a wide margin. On input modalities, d1-ultra takes text and images, whereas Classify-1 takes text, images and files, so the ScaleDown model accepts one extra kind of input. Both produce decisions as their only output modality. Neither model lists any supported parameters, so there is no difference to exploit there, and the qualitative comparison gives no meaningful difference on tool use, structured output or reasoning, so you should not assume one is stronger than the other on those axes. On price, the two are about the same for input, and the comparison gives nothing useful for output pricing, so treat them as comparable on cost rather than picking on price alone.

For practical work, the deciding factor is usually how much material you need the model to look at before it decides. If your task involves long documents, large batches of text, or anything that would blow past roughly thirty thousand tokens, ScaleDown: Classify-1 is the natural choice because of its much larger context window. If your inputs are short and you also want to hand the model an image alongside text, LiquidAI: d1-ultra covers that case, and it is the model to reach for when your pipeline is built around compact, image plus text decisions. When your inputs include files as a distinct input type rather than pasted text or images, Classify-1 is the one that accepts them directly.

Calling either model through api14 works the same way, because the gateway exposes an OpenAI-compatible API. You send your request to the same endpoint with the same request shape, and the only thing you change between the two is the model id: use liquid/d1-ultra for the LiquidAI model and scaledown/classify-1 for the ScaleDown model. Everything else in your client, your authentication and your response handling stays as it is, so switching between them is a one line change and you can compare their decisions on your own data without rewriting integration code.

There is one case where the smaller model is still the better pick even though the other has more context. If your inputs are short and your task is image plus text with no files involved, LiquidAI: d1-ultra fits the job and you do not need the extra capacity of Classify-1. Since the two are about the same price on input, there is no cost penalty for choosing the smaller context model when your inputs are small, and no reason to send more tokens than the task requires.

Switch with one string

Both models use the same endpoint and request shape. Change the model value from "liquid/d1-ultra" to "scaledown/classify-1" and prompts, streaming and tools keep working.

LiquidAI: d1-ultra details & playground →ScaleDown: Classify-1 details & playground →

import OpenAI from "openai";

const client = new OpenAI({
  baseURL: "https://api14.com/api/v1",
  apiKey: process.env.API14_API_KEY,
});

const result = await client.post("/decisions", { body: {
  "model": "scaledown/classify-1",
  "state": "Customer: I was charged twice for my order and want my money back.",
  "questions": {
    "team": {
      "type": "choice",
      "instructions": "Which team should handle this message?",
      "criteria": {
        "billing": "Payments, refunds, invoices",
        "support": "Product questions and bugs",
        "sales": "Pricing and new purchases"
      }
    },
    "urgent": {
      "type": "noul",
      "instructions": "Does this need a reply today?",
      "criteria": {
        "true": "Money, outage or legal risk",
        "false": "Can wait"
      }
    }
  }
} });
console.log(result.answers);

More comparisons

d1-ultra vs Classify-1: API price, context & features · api14