Claude Fable 5.2 Sets Intelligence Benchmark at 65, With Steep Cost Tradeoffs
Anthropic released Claude Fable 5.2 on November 9, 2026, achieving a score of 65 on the Artificial Analysis Intelligence Index—well above the median of 40 for comparable models. The model supports text and image inputs with a 1M-token context window and generates text output through extended reasoning capabilities.
The intelligence gains come with significant performance penalties. Claude Fable 5.2 costs $8 per million input tokens and $40 per million output tokens (median: $2 and $10 respectively), averaging $26.54 per evaluation task. Processing speed is particularly slow at 46.4 minutes per task versus a median of 12.7 minutes. The model also exhibits extreme verbosity, generating 275 million tokens during Intelligence Index evaluation compared to a median of 72.8 million—a substantial cost factor for users.
The model is proprietary with weights undisclosed, but available through six API providers. Its premium pricing and computational overhead position it for knowledge-intensive applications where accuracy justifies the resource consumption, rather than cost-sensitive deployments.
Read the full article
Continue reading at Hacker News →This is an AI-generated summary. Read the original for the full story.