6 September 2026
Artificial Analysis updates AI benchmark after criticism of GPT-6 Astra scoring
First reported
AI Breakfast ran this on , a day before the next source picked it up.
- Artificial Analysis released version 4.2 of its Intelligence Index, a ranking system that scores AI models. The update came after other evaluations ranked OpenAI's GPT-6 Astra much higher than Artificial Analysis initially did.
- With the new scoring, GPT-6 Astra gained four points and now ranks second overall, behind Anthropic's Claude Fable 5.1. Astra also uses fewer tokens (computational units) per task than competing frontier models.
How it was covered
Reported by The Decoder