Movers

Every benchmark lead that changed hands since the last edition, newest first. A check means a third party verified the score; the rest are lab-reported.

Every test where a different AI took the top spot recently. A check mark means someone outside the company confirmed the score.

Sectors: ten domains against human parity

How the labs define AGI

Register of releases, 2018 to now

Every model on the tracker, by year. Click a name for parameters, context, and the benchmarks it shipped with.

Every major AI model in order. Click a name to see its details.

Head-to-head

Every model, every column. The triangle marks the column leader; shading is position within the column's observed range. Click a header for the progression over time, a row to pin it. Blank cells mean no public figure, never an estimate.

How every model scores on the same tests. The triangle marks the best in each column. Click a test name to see how scores climbed over time.