Loading…
Loading…
ModelDex tracks the specs, pricing, benchmarks, and news for the leading AI models in one place, and ties every fact to a primary source and a verification date. When a number can't be confirmed against its source, it's flagged rather than published.
The frontier moves fast. Context windows change, prices get cut, new flagship models ship on a Tuesday. Keeping a clear, current picture of what each model actually does — and what it costs — is real work, and most references go stale the week after they publish.
ModelDex is built to stay current. Every published fact is tied to a primary source and a verification date, and anything that can't be confirmed against its source is flagged rather than served as fresh — so every number on the site can be traced back to where it came from.
Anyone learning the frontier models. Engineers picking an API, founders comparing pricing, analysts tracking the race, and people who simply want to understand the difference between one model and the next without wading through a dozen tabs.
ModelDex also powers Sentravision's own AI training work — when you teach people how these models differ, your source material has to be correct. The same reference Sentravision relies on internally is the one published here.
Every fact is bound to a primary source and a verification date. If we can't point to where a number came from, we don't publish it.
A correct, narrow dataset beats a broad one full of guesses. We'd rather mark a fact stale than serve it as if it were fresh.
The site is written to teach. Plain language, defined terms, and an honest view of what each model can and can't do.
ModelDex is a Sentravision project. It is an independent reference and is not affiliated with, endorsed by, or sponsored by OpenAI, Anthropic, Google, xAI, or any other model provider. Product names and trademarks belong to their respective owners. All facts are drawn from public, primary sources and are presented for reference only.