Schematron V2 Small

Inference.net launched Schematron V2 Small on September 12.It is not a general-purpose AI model. It does one thing: turn HTML into JSON.
Advertisement 728 × 90
CompanyInference Net
Context128K
Released2026-09
Updated2026-09-15

Schematron V2 Small Overview

Schematron V2 Small is an HTML-to-JSON extraction model from Inference.net.

It has 3B parameters and is fine-tuned from Llama 3.2 3B.

The way it works is different from a normal chat model.

At first, I assumed I could simply send a prompt like:

“Extract the product name, price, and stock information.”

That is not how Schematron works.

You define the extraction rules in a JSON Schema, put that schema inside the API request’s response_format, and the model returns JSON that matches the structure.

The benefit is predictability.

Field names, data types, and required values are fixed ahead of time. You spend less time cleaning up inconsistent outputs afterward.

The downside is flexibility.

If the webpage structure changes, you need to update the schema. You cannot just rewrite a prompt and expect the model to figure it out.

Schematron V2 Small supports a 128K token context window and up to 4096 output tokens per request.

It is mainly designed for:

web scraping
product data pipelines
financial document parsing
structured information extraction

The weights are not open source. Access is API-only.

Schematron V2 Small Pricing

PlanPriceDescription
API usage $0.05 / 1M input tokens; $0.23 / 1M output tokens Pay based on actual token usage

There is no subscription plan.

Inference.net does not currently offer an official free tier, although some platforms may provide trial credits.

The low price is one of its biggest advantages, but it only matters if your task matches what the model is built for.

Schematron V2 Small Key Features

1. Schema-driven extraction

The biggest difference between Schematron and general AI models is that it follows a schema instead of a prompt.

You do not ask:

“Find the product information.”

You define exactly what you want:

  • product name
  • price
  • category
  • attributes

The model then maps webpage content into those fields.

For large-scale data pipelines, this approach is much easier to manage than calling a general model and fixing the output afterward.

The limitation is that there is not much room for interpretation.

If the structure of the website changes, your schema needs to change too.

Schematron is better viewed as a structured extraction tool, not a webpage analyst.


2. Long-page handling

Schematron V2 Small supports a 128K token context window and is designed to handle long HTML pages.

Inference.net says the model is trained to deal with noisy webpage markup and messy real-world HTML.

That part matters more to me than a small benchmark improvement.

Real webpages are rarely clean. They contain ads, scripts, nested elements, and unnecessary content.

If a model can handle that without heavy preprocessing, it saves practical engineering time.

The company still recommends removing elements like scripts and styles with tools such as lxml.


3. Small model, low cost, and fast processing

Schematron V2 Small has only 3B parameters.

That is a major reason it can stay inexpensive.

Inference.net reports a throughput of 2.47 requests per second on a single H100, using a test setup with 10K input tokens and 500 output tokens.

It is not designed to compete on complex reasoning speed.

That is not the point.

Its job is repetitive processing: receive HTML, extract fields, return structured data.

Benchable testing places it in the 93rd percentile for speed, with 100% reliability in its reported evaluation.

I care less about the ranking itself and more about production behavior.

For an extraction model, a small number of formatting failures across thousands of pages can create more work than a slightly slower response.


4. V2 Small closes the gap with the larger V1 model

Schematron V1 had both 3B and 8B versions.

The goal of V2 Small is to get close to the 8B version while keeping the smaller model’s efficiency.

Inference.net used GPT-5.4 as an LLM judge:

  • V2 Small: 4.060
  • V1 8B: 4.070
  • V1 3B: 3.909

On SimpleQA:

  • V2 Small + GPT-5 Nano + Exa: 83.10
  • V1 8B: 85.58

The 8B model still scores slightly higher.

But I am not sure that difference matters much in production.

If a system is processing hundreds of thousands of webpages, cost, speed, and consistency usually matter more than a small benchmark gap.

That is where a smaller model can make more sense.

Summary

I would use Schematron V2 Small for:

  • product data extraction
  • webpage crawling pipelines
  • HTML-to-JSON conversion
  • structured document processing

These are tasks where I need reliable fields, not a model that explains what a page means.

I would not use it for:

  • content summaries
  • complex analysis
  • understanding the purpose of a webpage
  • open-ended questions

If the job is moving information from webpages into a database, Schematron makes sense.

If the job requires judgment, reasoning, or interpretation, I would use a general model instead.

Comments (0)

Leave a comment

Advertisement 728 × 90

Inference Net Model Comparison

Model Context Pricing API Released Global Heat
Schematron V2 Small
128K YES 2026-09
128K YES 2026-09

Similar Models

Schematron V2 Turbo
Inference.net released Schematron V2 Turbo on September 12, alongside V2 Small. Both models do the same job: converting HTML into JSON. The difference is the priority.V2 Small focuses more on extraction quality. Turbo focuses on throughput and cost.
Inference Net
Claude Opus 4.8
100
Claude Opus 4.8 is Anthropic's most powerful Opus-series model, featuring multimodal input, reasoning, and a 1M-token context window, excelling in complex reasoning and coding.
Anthropic
Claude Opus 5
100
What interests me most about Opus 5 isn’t the benchmark bump. It’s that Anthropic has pushed near-flagship capability into a price range more teams can actually use every day. I still wouldn’t treat it as a model you can leave alone for hours and assume everything will work out.
Anthropic
GPT-5.5 Pro
100
I’ve been using GPT-5.5 Pro with a few colleagues for the past several weeks, mostly on the kinds of jobs where regular chatbots tend to fall apart: long documents, messy research questions, code debugging, and tasks that need more than one or two steps of reasoning.
OpenAI
GPT-5.5
99
My first impression of GPT-5.5 is that it feels less like a chatbot and more like a coworker you can actually hand work to. Give it a task and it can break it down, use tools, and check what it has done.
OpenAI
GPT-5.6 Luna Pro
99
OpenAI’s next-generation efficient AI model, balancing speed, cost, and intelligence for chat, coding, and automation tasks.
OpenAI
DeepSeek V4.1 Flash
99
DeepSeek V4.1 Flash has taken over from Pro with a larger model, a 1-million-token context window, built-in vision, and lower API pricing. Some published scores already beat V4 Pro. What matters now is how that holds up in real workloads.
DeepSeek
Gemini 3.8 Flash
98
Gemini 3.8 Flash is a more interesting upgrade than its version number suggests. Google launched Gemini 3.7 Flash just three weeks ago, and now 3.8 Flash is already here. That pace is hard to ignore.
Google