GPT-6.2 Models Rise on AI Index, but Coding Scores Stall

Artificial Analysis put Astra second behind Claude Fable 5.2. Lower token prices improve the cost trade-off, while Sol lost three coding-index points.
This is the summary The Defiant publishes in its feed. The full article lives on their site.