UAB
Home/Data status
Data status
Live data status

What is current, and what needs attention.

The active catalog publication, independent source checks, retained last-known-good data, and factual copy reviews.
Published · Oct 3, 2026
Stale sources · 4
Claim warnings · 105
Active publication

Catalog snapshot

Published Oct 3, 2026Schema v213 open data reviews

Snapshot ID: catalog-v2-provider-models-20261003T114105Z

Source health

Latest passing data and latest attempt

SourceContentFreshnessLast captureLatest attemptWeight
ArenaverifiedrecentOct 3, 2026succeeded · Oct 3, 2026100%
LiveBenchverifiedrecentOct 3, 2026succeeded · Oct 3, 2026100%
Artificial AnalysisverifiedrecentOct 3, 2026succeeded · Oct 3, 2026100%
BridgeBenchverifiedarchivedMay 15, 2026failed · Oct 3, 202610%
Vals AIverifiedrecentOct 3, 2026failed · Oct 3, 2026100%
Terminal-BenchverifiedrecentOct 3, 2026succeeded · Oct 3, 2026100%
LLMBaserelayrecentOct 3, 2026succeeded · Oct 3, 2026100%
Scale LabsverifiedrecentOct 3, 2026succeeded · Oct 3, 2026100%
OpenCompassverifiedarchivedApr 22, 2026succeeded · Oct 3, 202610%
MTEBverifiedarchivedApr 22, 2026succeeded · Oct 3, 202610%
Provider official evalsverifiedarchivedApr 19, 2026succeeded · Oct 3, 202610%
Provider model registryverifiedrecentOct 3, 2026succeeded · Oct 3, 2026100%
Editorial accuracy

Claims due for review

105 warnings
BridgeBench: Coding-heavy evaluation focused on UI generation, debugging, refactoring, security, hallucination, and speed.

overdue · reviewed May 15, 2026 · every 30 days

Open surface
OpenCompass: Open evaluation platform with broad support for LLM and multimodal datasets.

overdue · reviewed Apr 22, 2026 · every 30 days

Open surface
MTEB: Embeddings and retrieval benchmark ecosystem for text and image across many languages and tasks.

overdue · reviewed Apr 22, 2026 · every 30 days

Open surface
Provider official evals: Manually verified benchmark results quoted from official provider model cards and release pages when third-party coverage lags behind the newest frontier models.

overdue · reviewed Undated · every 30 days

Open surface
Code Arena: Legacy identifier for WebDev Arena measurements, retained for historical source rows. Current measurements use WebDev Arena.

overdue · reviewed Undated · every 30 days

Open surface
Search Arena: Blind preference arena for search-grounded answers and web-connected browsing behavior.

overdue · reviewed Aug 24, 2026 · every 30 days

Open surface
Video Edit Arena: Blind preference arena for instruction-based video editing quality.

overdue · reviewed Aug 27, 2026 · every 30 days

Open surface
Vision Arena · Creative Writing: Style-controlled Vision Arena creative writing leaderboard from the public Arena dataset.

overdue · reviewed Jan 9, 2026 · every 30 days

Open surface
Vision Arena · Creative Writing · No Style Control: Vision Arena creative writing leaderboard without style-control adjustment from the public Arena dataset.

overdue · reviewed Jan 9, 2026 · every 30 days

Open surface
Text-to-Image Arena · Chinese: Text-to-Image Arena chinese leaderboard from the public Arena dataset.

overdue · reviewed Jan 24, 2025 · every 30 days

Open surface
Text-to-Image Arena · English: Text-to-Image Arena english leaderboard from the public Arena dataset.

overdue · reviewed Feb 11, 2025 · every 30 days

Open surface
Text-to-Image Arena · Russian: Text-to-Image Arena russian leaderboard from the public Arena dataset.

overdue · reviewed Jan 24, 2025 · every 30 days

Open surface
Search Arena · No Style Control: Search Arena overall leaderboard without style-control adjustment from the public Arena dataset.

overdue · reviewed Aug 24, 2026 · every 30 days

Open surface
Overall: LiveBench global score averaged across the current website category scores.

overdue · reviewed Jun 25, 2026 · every 30 days

Open surface
Reasoning: Recent objective reasoning tasks from the current LiveBench website release.

overdue · reviewed Jun 25, 2026 · every 30 days

Open surface
Coding: LiveBench code generation and completion score from the current website release.

overdue · reviewed Jun 25, 2026 · every 30 days

Open surface
Agentic coding: LiveBench agentic coding tasks spanning JavaScript, TypeScript, and Python.

overdue · reviewed Jun 25, 2026 · every 30 days

Open surface
Mathematics: LiveBench math category across competition, olympiad, integral, and AMPS-hard tasks.

overdue · reviewed Jun 25, 2026 · every 30 days

Open surface
Data analysis: LiveBench data-analysis category across event, table-join, and table-reformat tasks.

overdue · reviewed Jun 25, 2026 · every 30 days

Open surface
Language: LiveBench language category across connections, plot unscrambling, and typo repair.

overdue · reviewed Jun 25, 2026 · every 30 days

Open surface
Instruction following: LiveBench instruction-following category across paraphrase, simplify, story, and summarize tasks.

overdue · reviewed Jun 25, 2026 · every 30 days

Open surface
Theory of mind: Theory of mind task slice from the current LiveBench website leaderboard.

overdue · reviewed Jun 25, 2026 · every 30 days

Open surface
Zebra puzzle: Zebra puzzle task slice from the current LiveBench website leaderboard.

overdue · reviewed Jun 25, 2026 · every 30 days

Open surface
Spatial: Spatial task slice from the current LiveBench website leaderboard.

overdue · reviewed Jun 25, 2026 · every 30 days

Open surface
Logic with navigation: Logic with navigation task slice from the current LiveBench website leaderboard.

overdue · reviewed Jun 25, 2026 · every 30 days

Open surface
Coding generation: Coding generation task slice from the current LiveBench website leaderboard.

overdue · reviewed Jun 25, 2026 · every 30 days

Open surface
Coding completion: Coding completion task slice from the current LiveBench website leaderboard.

overdue · reviewed Jun 25, 2026 · every 30 days

Open surface
JavaScript: JavaScript task slice from the current LiveBench website leaderboard.

overdue · reviewed Jun 25, 2026 · every 30 days

Open surface
TypeScript: TypeScript task slice from the current LiveBench website leaderboard.

overdue · reviewed Jun 25, 2026 · every 30 days

Open surface
Python: Python task slice from the current LiveBench website leaderboard.

overdue · reviewed Jun 25, 2026 · every 30 days

Open surface