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.
Inference.net: Schematron V2 Small
inference-net/schematron-v2-smallTry Inference.net: Schematron V2 Small now
Runs on your balance at the prices above. New accounts start with free usage.
One API. Every model.
Session cost $0.00
Settings
Advanced parameters
Inference.net: Schematron V2 Small at a glance
- Input price is 89% lower than the median of 434 Text models.
- Context window of 128,000 tokens, enough for long documents and multi-turn conversations.
- Supports structured JSON output with a schema.
- A typical request (1,000 input / 500 output tokens) costs about $0.000578.
About Inference.net: Schematron V2 Small
Inference.net: Schematron V2 Small is a text model from Inference.net. It is a 3B-parameter HTML-to-JSON extraction model, and the provider describes it as prioritising extraction quality for complex schemas and long pages. It sits in the text category, so it is aimed at language and document processing work rather than images, audio or other modalities.
The model takes text as input and returns text as output. Its context window is 128,000 tokens, which is what allows it to work through long pages in a single request, and its maximum output is 4,096 tokens. The provider notes that extraction instructions must be supplied through a JSON schema, so the schema is the mechanism that tells the model what to pull out of the HTML and how to shape the result.
Those characteristics point to a fairly specific set of uses. The obvious one is turning HTML pages into structured JSON, for example pulling product details, listings, article metadata or other repeated fields out of pages that vary in layout. Because the model is built for complex schemas, it suits cases where the target structure has nested objects or many fields rather than a couple of flat values. The long context helps when the relevant content is buried far down a page or when a single page contains many records to extract. It is a text-in, text-out model, so it fits pipelines where HTML is fetched elsewhere and handed to the model for parsing.
You call it through api14's OpenAI-compatible API using the model id "inference-net/schematron-v2-small". The request parameters it supports are frequency_penalty, logit_bias, max_tokens, min_p, presence_penalty, repetition_penalty, response_format, seed, stop, structured_outputs, temperature, top_k and top_p. For extraction work, response_format and structured_outputs are the parameters most directly relevant to getting JSON back in the shape you asked for, while seed helps when you want repeatable results across runs. Sampling controls such as temperature, top_p, top_k and min_p let you trade variation for consistency, and max_tokens caps the length of the returned JSON.
Keep the output limit in mind: 4,096 tokens is the ceiling on a single response, so very large extraction jobs may need to be split across multiple calls even though the input side can hold 128,000 tokens. The model is also text-only, so pages that depend on images or rendered layout will need to be handled before they reach it. If your task is general chat, reasoning or generation rather than schema-driven extraction from HTML, a general purpose text model is likely a better fit.
How to call the Inference.net: Schematron V2 Small API
- Create an API key in your dashboard.
- Install the official openai SDK for your language.
- Set baseURL to https://api14.com/api/v1 and model to "inference-net/schematron-v2-small".
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": "inference-net/schematron-v2-small",
"messages": [
{
"role": "user",
"content": "Explain vector databases in two sentences."
}
]
});
console.log(result.choices[0].message.content);
Supported parameters
frequency_penaltylogit_biasmax_tokensmin_ppresence_penaltyrepetition_penaltyresponse_formatseedstopstructured_outputstemperaturetop_ktop_pSpecifications
- Context length
- 128,000
- Max output
- 4,096
- Input
- text
- Output
- text
Requests on api14 · last 7 days
Inference.net: Schematron V2 Small: frequently asked questions
- How much does the Inference.net: Schematron V2 Small API cost?
- Inference.net: Schematron V2 Small costs Input $0.18 / 1M tokens. It is charged from your USD balance only for successful requests; failed requests are free.
- What is the context window of Inference.net: Schematron V2 Small?
- Inference.net: Schematron V2 Small reads up to 128,000 tokens per request: prompt, history and documents together. It writes up to 4,096 output tokens per response.
- What can Inference.net: Schematron V2 Small do?
- Inference.net: Schematron V2 Small accepts text and returns text. It supports structured JSON output.
- How do I call Inference.net: Schematron V2 Small from my code?
- Point any OpenAI SDK at https://api14.com/api/v1 and set the model to "inference-net/schematron-v2-small". No other code changes. Copy a ready example from the code panel on this page.
- What are alternatives to Inference.net: Schematron V2 Small?
- Comparable Text models at a similar price: Mistral: Mistral Small 3.2 24B, DeepSeek: DeepSeek V4 Flash 0423, Poolside: Laguna S 2.1. Each runs through the same API key, so switching is one string.
Inference.net: Schematron V2 Small community
New topicThis community has no topics yet. Open the first one about Inference.net: Schematron V2 Small.