Ridiculous benchmarks.

Very precise measurements. Deeply questionable priorities.

Felony Bench

Source-reported agent incidents, with the reports behind each count.

The version arms race. Six beats five. Obviously.

Every character counts. Even the spaces are pulling their weight.

The only leaderboard where Q is worth ten points.

One generation. Several outfits. Who's under the coat?

Bigger keyboard journeys. Bigger imaginary brains.

Launch on the 31st. Bring every variant. Collect your points.