Artificial Analysis: Haiku 5.5's highest reasoning tier scores 43 on the Intelligence Index

Anthropic released Claude Haiku 5.5, scoring 43 on the Artificial Analysis Intelligence Index, a 26-point improvement over the previous Haiku a year after its release. This is the first Haiku model to support effort settings and adaptive thinking, and introduces tiered pricing: $0.10/$0.50 per 1M input/output tokens for requests under 100k tokens, rising to $0.50/$2.50 above 100k. Its context window increased from 200k on Claude 4.5 Haiku to 1M tokens, and Terminal-Bench 4.0 scores rose from 0% to 33%, but the AutomationBench-AA score of 35% may be underestimated due to safety refusal issues.