Sources / arc-prize

ARC Prize Leaderboard

healthydefault: independent

arcprize.org/leaderboard · trust rank 80 · adapter arc-prize

Observations

1,226

append-only score facts

Raw snapshots

immutable fetched payloads

Last success

2026-09-12 05:30Z

end of last succeeded/partial run

Freshness

2.8 h

hours since last success

License & attribution

Task datasets are Apache-2.0 and the harness/raw results MIT; the leaderboard JSONs are unlicensed with no access prohibition. Cited per ARC Prize policy (arcprize.org/policy).

What we ingest

Static leaderboard JSONs (evaluations, models, datasets): per-configuration accuracy and cost-per-task across ARC-AGI-1/2/3. Semi-private and public eval splits never mix — the semi-private split is the canonical variant, and cost never enters the composite.

Ingestion runs

last 10 of the run history

Started	Duration	Status	Key stats
2026-09-12 05:30Z	8.9 s	partial	docs 3 · staged 1,823 · promoted 1,300 · blocked 16,114 · scoreAdded 0 · unchanged 3
2026-09-11 05:30Z	13.0 s	partial	docs 3 · staged 1,823 · promoted 1,300 · blocked 15,591 · scoreAdded 0 · unchanged 3
2026-09-10 05:30Z	17.9 s	partial	docs 3 · staged 1,823 · promoted 1,300 · blocked 15,068 · scoreAdded 0 · unchanged 3
2026-09-09 05:30Z	8.1 s	partial	docs 3 · staged 1,823 · promoted 1,300 · blocked 14,545 · scoreAdded 0 · unchanged 3
2026-09-08 05:30Z	10.4 s	partial	docs 3 · staged 1,823 · promoted 1,300 · blocked 14,022 · scoreAdded 0 · unchanged 3
2026-09-07 05:30Z	9.8 s	partial	docs 3 · staged 1,823 · promoted 1,300 · blocked 13,499 · scoreAdded 0 · unchanged 3
2026-09-06 05:30Z	7.6 s	partial	docs 3 · staged 1,823 · promoted 1,300 · blocked 12,976 · scoreAdded 0 · unchanged 3
2026-09-05 05:30Z	11.6 s	partial	docs 3 · staged 1,823 · promoted 1,300 · blocked 12,453 · scoreAdded 0 · unchanged 3
2026-09-04 05:30Z	11.0 s	partial	docs 3 · staged 1,823 · promoted 1,300 · blocked 11,930 · scoreAdded 0 · unchanged 1
2026-09-03 05:30Z	23.7 s	partial	docs 3 · staged 1,756 · promoted 1,300 · blocked 11,407 · scoreAdded 0 · unchanged 3

partial = run finished but some rows are blocked on identity review — nothing is published until every referenced model/agent/benchmark is resolved.

Raw snapshots

latest 5

Fetched	URL	SHA-256	Last re-fetch
2026-09-04 05:30Z	https://arcprize.org/media/data/models.json	2a3ac802eeda…	2026-09-12 05:30Z
2026-09-04 05:30Z	https://arcprize.org/media/data/evaluations.json	17b35f029522…	2026-09-12 05:30Z
2026-09-02 05:30Z	https://arcprize.org/media/data/models.json	95b03912ddee…	2026-09-03 05:30Z
2026-09-02 05:30Z	https://arcprize.org/media/data/evaluations.json	d20f203e9811…	2026-09-03 05:30Z
2026-08-22 05:30Z	https://arcprize.org/media/data/models.json	17aeed69df45…	2026-09-01 05:30Z

Snapshots are immutable raw payloads — every published number traces back to one. Re-fetching an identical payload bumps the re-fetch time instead of storing a duplicate. Scores from this source appear on benchmark profiles and model pages with this source linked.