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

2026-07-14 00:15Z

end of last succeeded/partial run

Freshness

1.4 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 3 of the run history

Started	Duration	Status	Key stats
2026-07-14 00:15Z	0.5 s	succeeded	docs 3 · promoted 0 · blocked 0 · scoreAdded 0 · unchanged 3
2026-07-14 00:08Z	0.3 s	succeeded	docs 3 · promoted 0 · blocked 0 · scoreAdded 0 · unchanged 3
2026-07-13 20:03Z	0.8 s	partial	docs 3 · staged 1,300 · promoted 0 · blocked 1,300 · scoreAdded 0 · unchanged 0

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 3

Fetched	URL	SHA-256	Last re-fetch
2026-07-13 20:03Z	https://arcprize.org/media/data/datasets.json	ad1595f5e707…	2026-07-14 00:15Z
2026-07-13 20:03Z	https://arcprize.org/media/data/models.json	ce23f13ca83c…	2026-07-14 00:15Z
2026-07-13 20:03Z	https://arcprize.org/media/data/evaluations.json	40b5be775302…	2026-07-14 00:15Z

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.