Anthropic Releases Claude Fable 5.2, Tops Artificial Analysis Intelligence Index Anthropic released Claude Fable 5.2 and Claude Mythos 5.2 early this morning, claiming improvements in programming, inference, and vision. Data from third-party benchmarking firm Artificial Analysis shows that Fable 5.2 scored 65 points at maximum inference intensity, ranking first. It is 5 points higher than the second-place Google Gemini 4 Carbon, 6 points higher than its own Opus 5.5, and approximately 15 points higher than the previous generation Fable 5.1. In terms of price, Fable 5.2's API cost is $8 per million tokens input and $40 per million tokens output, 20% cheaper than the previous generation. However, it consumes more tokens at maximum inference intensity, averaging $26.50 per task, roughly twice that of Fable 5.1.