Claude Fable 5.2 Sets Record With 65 on Artificial Analysis Intelligence Index

At a glance

Anthropic's AI model Claude Fable 5.2 scored 65 on the Artificial Analysis Intelligence Index, well above the 40-point median for models in a similar price range. However, it took much more time and money to complete a task than other models.

Key points

1. 1

   The estimated time to complete one task was 44.1 minutes, about 3.4 times the comparison models' median of 12.9 minutes.

2. 2

   The API price is $8 per million input tokens and $40 per million output tokens, and the average cost per evaluation run was $26.54.

3. 3

   Across the evaluation, it generated 275 million output tokens, about 3.8 times the comparison models' median.

4. 4

   It is a reasoning model that takes image and text input and responds with text, and it can handle 1 million tokens of context at once.

Why it matters

It shows that even as AI model performance improves, how much time and money it takes to actually use the model matters just as much.