About the product

Modelglass is a structured pricing and capability registry for AI models — tracked across LLM, image, video, audio, coding, science, and agentic benchmarks — built for developers and AI-native teams who need to query, compare, and route against real, versioned data instead of scattered vendor docs and stale blog posts.

Every entry starts as a YAML file: pricing, benchmark scores, context windows, modality support. Each file is schema-validated and cross-referenced against benchmark definitions, then compiled into a single content-hashed JSON artifact per vertical. That artifact is the source of truth for everything downstream — the web app, a three-tier REST API, an MCP server you can point an agent at directly, a VS Code extension for cost-aware routing, and comparison/timeline views for tracking how a model's specs shift over time.

The registries update daily via automated agents that propose changes — nothing merges without a human reviewing it first. If a price or benchmark looks stale, it's flagged, not guessed.

If you're building something that needs to pick the right model for the job — or the cheapest one that's good enough — that's what this exists for.

API
MCP
VS Code
iOS app

Registry

Verticals and benchmarks · verified attributes and pricing · updated daily


About Modelglass

Modelglass was built by Scott Schinkel, a data professional with over 20 years of experience turning complex systems into legible, actionable information.

The idea behind Modelglass is simple: AI models — for image, language, video, and audio generation, and for coding, science, and agentic capability — have become genuinely powerful, but comparing them across providers is needlessly difficult. Pricing is fragmented, capability claims are inconsistent, and there's no neutral place to see the full picture at once.

Modelglass exists to fix that — one source of truth for model pricing, capability profiles, and benchmarks across every modality and every capability vertical, kept current and presented without bias.

Scott believes that good decisions come from good data. That means data that's complete, honestly presented, and structured so the signal is easy to find — whether you're routing inference at scale or just trying to figure out which model is worth your time.

The idea for Modelglass came out of a company All Hands in May 2026, listening to executives describe travelling between countries just to understand the differences between AI models and providers — a landscape adding products and differentiators faster than anyone can track by hand. I wanted a level playing field: something people could trust and rely on for AI model selection, sourced daily rather than guessed from stale blog posts.

My data story goes back further than this project. I started in the Australian financial industry, in superannuation, building early data products to understand competitors, alongside heavily regulated financial reporting work. For the last 11 years, I worked at a Melbourne tech company across many different ingestion patterns and warehouse types — always focused on serving data that had utility, not just data that existed.

Modelglass started humbly. When it went live in June 2026, it tracked models across image, language, video, and audio. Since then the registry has more than doubled to over 160 models, with benchmark tracking on top. It's available via API and MCP server, alongside a set of VS Code routing tools — all built around the same idea: a data product people can trust, in an environment where we're not trying to control the assets being generated.

Read the full founder story on Medium →

Follow on X →