Haiku 5.5 puts cheaper specialist agents in the spotlight

Anthropic released Claude Haiku 5.5 on October 7. For prompts up to 100,000 tokens, listed API prices are $0.10 per million input tokens and $0.50 per million output tokens; larger prompts use higher rates. The model adds adjustable effort and is positioned for summaries, classification, browser work, and coding subagents. Anthropic also cut Sonnet 5.5 cache-read pricing and announced monthly API credits for eligible Max and Team subscribers.

The useful comparison is the cost of an accepted result, including retries and review. Routing narrow tasks to a smaller model could improve economics, while difficult repository changes still need stronger reasoning and validation. Anthropic explicitly positions Sonnet and Opus above Haiku for complex agentic coding. Roseram has not independently reproduced the launch benchmarks.