Changelog

What changed

Every change to the dataset, newest first — new and removed models, tools and benchmarks, and anything updated in between. Record-level only; see each record's page for the current detail.

Finish Phases 2 and 3: three new tasks, per-lane tool boards, fetchers, logos (D20–D22)

60 updated models · 42 new models · 33 new tools · 39 updated tools · 14 new benchmarks · 9 updated benchmarks · 3 new tasks · 4 updated tasks

Verify flagged records and stale boards (D19)

1 new model · 23 updated models · 1 removed tool · 22 updated benchmarks · 1 new benchmark · 2 removed benchmarks · 2 updated tasks

Discovery pass: long-tail tools and open models for all 24 tasks (D18)

197 new models · 57 updated models · 384 new tools · 48 updated tools · 6 updated benchmarks · 8 new benchmarks · 10 updated tasks

Expand to 24 tasks in 7 groups, add the task overview and positioning preview

136 new models · 53 updated models · 162 new tools · 41 updated tools · 39 new benchmarks · 2 updated benchmarks · 20 new tasks · 4 updated tasks