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.

Extract vs Xing4.0 29B A4B (free): API price, context & features · api14

ScaleDown: Extract vs XingChen: Xing4.0 29B A4B (free)

A side-by-side comparison of ScaleDown: Extract ($0.14 / 1M tokens) and XingChen: Xing4.0 29B A4B (free) (Free) using current USD prices and capabilities from the api14 catalog. Both run through the same OpenAI-compatible API.

Specifications

	ScaleDown: Extract	XingChen: Xing4.0 29B A4B (free)
Provider	ScaleDown	XingChen
Input	$0.14 / 1M tokens	-
Context window	1,000,000	262,144
Max output	900,000	235,929
Input	text, image, file	text
Output	text	text
Tool calling
Structured output
Reasoning control
Added	Oct 13, 2026	Oct 16, 2026

Key differences

ScaleDown: Extract reads 3.8× more context (1,000,000 tokens).
Only XingChen: Xing4.0 29B A4B (free) supports tool calling.
Only XingChen: Xing4.0 29B A4B (free) supports structured JSON output.
Only XingChen: Xing4.0 29B A4B (free) supports reasoning effort control.

In depth

ScaleDown: Extract is a text model from ScaleDown, released on 2026-10-13. XingChen: Xing4.0 29B A4B (free) is a text model from XingChen, released on 2026-10-16. Both are served through api14 as OpenAI-compatible chat models, and both are listed in the text category. The naming tells you something about positioning: Extract sounds like a purpose-built extraction model, while Xing4.0 29B A4B is a general chat and reasoning model with a free tier.

The two differ most in context and modalities. ScaleDown: Extract has a context window of 1,000,000 tokens, which is the larger of the two, and it accepts text, image and file inputs, producing text output. XingChen: Xing4.0 29B A4B has a 262,144 token context window and accepts text only. On parameters the split runs the other way. XingChen exposes a long list of sampling and control parameters: frequency_penalty, presence_penalty, temperature, top_p, top_k, min_p, repetition_penalty, seed, stop, max_tokens, response_format, structured_outputs, tool_choice, tools, reasoning and include_reasoning. ScaleDown: Extract supports only max_tokens. So XingChen is the only one of the two with tool calling, structured outputs and reasoning, while ScaleDown is the only one that takes images and files and the only one with the larger context.

For practical work, the choice follows directly from those differences. If your task involves long documents, image or file inputs, or a single prompt that has to hold a very large amount of material, ScaleDown: Extract is the natural pick because of its larger context and its multimodal input support. If your task needs function calling, JSON-schema-constrained output, or a visible reasoning pass, XingChen: Xing4.0 29B A4B is the pick, since those are capabilities ScaleDown: Extract does not expose. For plain text generation where none of those features matter, either will do, and the live price table on this page is the better guide than anything written here.

Calling either model through api14 works the same way, because the endpoint is OpenAI-compatible. You point your client at the api14 base URL, send a chat completions request, and set the model field to the id you want. To switch models you change only that string, from "scaledown/extract" to "xingchen-agi/xing4.0-29b-a4b:free" or back. Your request body, authentication header and response parsing stay as they are. The one caveat is parameters: if you send tools, response_format or reasoning fields, they apply to XingChen and are not supported by ScaleDown: Extract, which accepts only max_tokens, so keep per-model parameter sets rather than one shared payload.

There are cases where the other model is still the better pick even after that. If you are already using XingChen for a tool-calling or structured-output pipeline and you only need a short text answer, staying on XingChen avoids a second integration. Conversely, if you are on ScaleDown: Extract for document work and you need a quick plain-text completion, staying there avoids adding another model id to your config. Neither model is strictly better; the deciding factors are context size, multimodal input, and whether you need tools, structured output or reasoning.

Switch with one string

Both models use the same endpoint and request shape. Change the model value from "scaledown/extract" to "xingchen-agi/xing4.0-29b-a4b:free" and prompts, streaming and tools keep working.

ScaleDown: Extract details & playground →XingChen: Xing4.0 29B A4B (free) 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.chat.completions.create({
  "model": "xingchen-agi/xing4.0-29b-a4b:free",
  "messages": [
    {
      "role": "user",
      "content": "Explain vector databases in two sentences."
    }
  ]
});
console.log(result.choices[0].message.content);

ScaleDown: Extract vs XingChen: Xing4.0 29B A4B (free)

Specifications

Key differences

In depth

Switch with one string

More comparisons