Fable 5.2 scores 65 in testing but takes 45.0 minutes per task

Artificial Analysis measured Claude Fable 5.2 at 65 on its Intelligence Index, against a median of 40 for comparable models, but found substantial cost and latency tradeoffs. The tested Max, Default Fallback configuration averaged $26.54 and 45.0 minutes per task; comparable models had a median task time of 12.9 minutes. The model generated 275M output tokens during the evaluation, compared with a median of 72.8M. It supports text and image input, text output and a 1M-token context window. The results apply to this specific reasoning configuration.