Member-only story
I am not going to hold you hostage. I prefer GPT-5.6 Sol, by a wide margin. If you have the time, let me explain why.
Claude Fable 5 is still a serious model. Artificial Analysis gives it 60 on the broad Intelligence Index, one point ahead of Sol at 59. It also has the stronger AA-Briefcase result for analytical knowledge work.
That still does not make Fable my default. Artificial Analysis puts Sol at $1.04 per Intelligence Index task, about one third of Fable 5’s estimated task cost. Sol also leads its Coding Agent Index at 80. Once I am choosing a model for repeated browser work, code changes, retries, tool calls, and long-running automation, that is the tradeoff I care about.
I do not want to pay three times as much for one point #
The public Intelligence Index picture is straightforward: Fable 5 scores 60, Sol scores 59.
One point is real. I am not pretending it does not exist. Fable also leads AA-Briefcase overall, with a 56% Rubric Score against Sol’s 42% and an Analytical Quality Elo of 1764 against Sol’s 1592. I…