Sources / scale-labs

Scale Labs

healthydefault: independent

labs.scale.com/leaderboard · trust rank 70 · adapter scale-labs

Observations

262

append-only score facts

Raw snapshots

313

immutable fetched payloads

Last success

2026-09-11 10:30Z

end of last succeeded/partial run

Freshness

21.8 h

hours since last success

License & attribution

scale.com terms grant an internal-purposes license and robots.txt allows the leaderboard, but republication is a legal gray zone — a wave-2 source, attributed and linked, to be reassessed before public launch.

What we ingest

Embedded payloads from labs.scale.com leaderboard pages: score plus 95% CI half-width per model configuration across their SEAL-lineage benchmarks (HLE, MultiChallenge, Fortress, …). Scale runs the evals, so rows land as independent.

Ingestion runs

last 10 of the run history

Started	Duration	Status	Key stats
2026-09-11 10:30Z	5.5 s	partial	docs 5 · staged 218 · promoted 215 · blocked 12 · scoreAdded 0 · unchanged 0
2026-09-10 10:30Z	5.1 s	partial	docs 5 · staged 218 · promoted 215 · blocked 9 · scoreAdded 0 · unchanged 0
2026-09-09 10:30Z	5.0 s	partial	docs 5 · staged 216 · promoted 215 · blocked 6 · scoreAdded 0 · unchanged 0
2026-09-08 10:30Z	4.5 s	partial	docs 5 · staged 216 · promoted 215 · blocked 5 · scoreAdded 0 · unchanged 0
2026-09-07 10:30Z	4.6 s	partial	docs 5 · staged 216 · promoted 215 · blocked 4 · scoreAdded 0 · unchanged 0
2026-09-06 10:30Z	5.3 s	partial	docs 5 · staged 216 · promoted 215 · blocked 3 · scoreAdded 0 · unchanged 0
2026-09-05 10:30Z	5.5 s	partial	docs 5 · staged 216 · promoted 215 · blocked 2 · scoreAdded 0 · unchanged 0
2026-09-04 10:30Z	4.6 s	partial	docs 5 · staged 216 · promoted 215 · blocked 1 · scoreAdded 0 · unchanged 0
2026-09-03 10:30Z	4.5 s	succeeded	docs 5 · staged 215 · promoted 215 · blocked 0 · scoreAdded 0 · unchanged 0
2026-09-02 10:30Z	4.7 s	succeeded	docs 5 · staged 215 · promoted 215 · blocked 0 · 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 5

Fetched	URL	SHA-256	Last re-fetch
2026-09-11 10:30Z	https://labs.scale.com/leaderboard/visual_language_understanding	8286387e6b3d…	2026-09-11 10:30Z
2026-09-11 10:30Z	https://labs.scale.com/leaderboard/mask	a3a261f32c7b…	2026-09-11 10:30Z
2026-09-11 10:30Z	https://labs.scale.com/leaderboard/enigma_eval	578d120bd254…	2026-09-11 10:30Z
2026-09-11 10:30Z	https://labs.scale.com/leaderboard/multichallenge	eb1aa57cadd4…	2026-09-11 10:30Z
2026-09-11 10:30Z	https://labs.scale.com/leaderboard/humanitys_last_exam	a09255106d50…	2026-09-11 10:30Z

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.