Artificial Analysis: GPT-5.6 Sol Lands at 59 on the Intelligence Index — One Point Under Claude Fable 5 at Roughly a Third the Cost per Task
Artificial Analysis·medium signal
Artificial Analysis benchmarks GPT-5.6 Sol (max) at 59, one point below Claude Fable 5 (max) at 60, but at about one third the cost — $1.04 per Intelligence Index task. Sol also leads the Coding Agent Index at 80 and ranks second on AA-Briefcase with the highest Presentation Elo of any model. Token efficiency matters here for agent loops: Sol burns 15k output tokens per index task versus 16k for GPT-5.5, fewer than most models of comparable intelligence. Terra and Luna score 55 and 51 at 50-80% lower cost.