Calm over noise
Understanding what changed shouldn't require watching everything all day. A record you can return to beats a feed you have to keep up with.
About
A calm, source-backed record of AI model change.
Modelmark helps people understand what changed across AI models, where the information came from, and why it may matter — without the daily chase.
Why we're building it
The pace is fine. The record is missing.
AI models change constantly — releases, updates, access shifts, pricing moves, capability signals. The information exists, but it's scattered, loud, and often stripped of its sources by the time it reaches you.
Modelmark is building the missing layer: a structured, reviewed, source-backed record of that change. Not a news feed, and not a ranking — a reference you can trust and check.
Modelmark tracks public AI labs and model records. It is independent, and not affiliated with the labs it tracks unless explicitly stated. Trust comes from sources, confidence labels, timestamps, and visible corrections — not from a brand voice.
One record, two surfaces
The app is for people. The API and MCP are for systems.
The app is the human-facing surface. The API and MCP are the system-facing surfaces. The record is the same. The provenance travels with it.
The iOS app is the human-facing surface today. The same source-backed record is being prepared to support API and MCP access for builders and agents.
What we believe
Three commitments that shape every record.
Sources over claims
Every record points back to where it came from. Evidence is cited, not paraphrased into something that sounds confident.
Honesty over hype
We say what we don't know, correct in the open, and avoid claiming capabilities — our own included — before they exist.
Where things stand
Modelmark is preparing private beta infrastructure. This website explains the methodology and direction; it does not represent a fully public production data product yet.