Modelglass is the pricing and capability intelligence platform for AI model selection. We track live pricing, benchmark performance, and capability ratings across 100+ language, image, video, and audio models. We give developers and operators the data they need to make confident, cost-efficient decisions.

The AI model landscape changes constantly. New models launch, prices shift, and the right choice for one workload is the wrong choice for another. Modelglass keeps pace with that change: every pricing entry is sourced, verified, and timestamped; every capability rating is backed by published benchmark data with transparent provenance.

That data now goes to work at the point of execution. The Modelglass VS Code extension routes a development task to the cheapest model that clears a real benchmark bar for that task type, then executes it directly against your own provider key, fully client-side, no proxy in the request path. It's the same registry data that powers the site and API, doing real work in your editor.

Beyond routing, Modelglass helps teams understand which models are genuinely strong at coding, tool use, or instruction following, and which are the cost-efficient choice for high-volume or low-complexity tasks, less overspending on premium models for work that doesn't need them, better outcomes when it does.

Modelglass is available via a web platform, a REST API, an MCP server, and the Modelglass Cost-Aware Router for VS Code.

Modelglass

Org chart

Teams

Offices

Modelglass

Org chart

Teams

Offices

Related companies