Reported evidence
Claude Haiku 5.5 → compares → Claude Haiku 4.5
AWS relays Anthropic’s claim that Claude Haiku 5.5 costs around 75% less than Claude Haiku 4.5 for most tasks. This is a reported vendor comparison, not an independently verified reduction in per-call prices or enterprise total costs; the excerpt does not specify task composition, quality constraints, or calculation methods.
Read the evidence
AWS announced the availability of Claude Haiku 5.5 and relayed Anthropic’s statement: the model is intended for subagents and high-volume, cost-sensitive tasks, with costs around 75% lower than Haiku 4.5 for most tasks.[S7]
[S7] AWS Machine Learning Blog
Read the full research