Astra and Fable split leads in Artificial Analysis Max-setting tests
The Defiant reports that in Artificial Analysis' Max-setting tests, Astra leads on GDP.pdf while Fable leads a factual-question net score that penalizes wrong answers. The post states that neither model establishes a winner for live research or reliable citations.
Why “Unverified”
- Reported by The Defiant alone, with no second source or primary confirmation yet
1 independent source group across 1 post · evidence score 52/100. How labels work
What the sources claim
-
In Artificial Analysis' Max-setting tests, Astra leads on GDP.pdf.
“In Artificial Analysis’ Max-setting tests, Astra leads GDP.pdf”
The Defiant -
Fable leads a factual-question net score that penalizes wrong answers.
“Fable leads a factual-question net score that penalizes wrong answers”
The Defiant -
Neither model establishes a winner for live research or reliable citations.
“Neither establishes a winner for live research or reliable citations”
The Defiant
Sources
More in AI & Tech
Embed this page
Summary and claims extracted by AI with prompt analyze-story@5 on 2026-11-18 18:30 UTC. Labels are computed from the sources, not by the model.