AI news · Anthropic · models
Anthropic cuts small-model prices with Claude Haiku 5.5
The new small model matches GPT-6 Luna’s per-token price for prompts up to 100,000 tokens and adds effort settings.
Anthropic released Claude Haiku 5.5 on Wednesday 7 October 2026, a small model for high-volume work such as summaries, classification and database queries. The company says it is its fastest model so far and the first Haiku with an adjustable effort setting, and that it costs about 75% less to run on average than Haiku 4.5.
List prices fall harder. For prompts up to 100,000 tokens, Haiku 5.5 costs $0.10 per million input tokens and $0.50 per million output, a tenth of Haiku 4.5’s rate, SiliconANGLE reports. That matches GPT-6 Luna. On Anthropic’s figures, Haiku 5.5 scores 72.4% on the OSWorld 2.1 offline subset against 48.9% for Luna. Anthropic also halved Sonnet 5.5 cache reads to $0.10 per million tokens.
Simon Willison found the catches. Above 100,000 tokens the price rises fivefold, to $0.50 and $2.50, while Luna’s step comes later and smaller. A new tokenizer counted about 1.25 times as many tokens on the same long prompt in his trial, a quiet increase.
For teams running classification, tagging or support bots at volume, Haiku 5.5 matches Luna’s price and posts higher benchmark scores up to 100,000 tokens; above that, Willison finds Luna the better deal.
Sources
- Anthropic: Introducing Claude Haiku 5.5. Checked
- SiliconANGLE: Anthropic releases Claude Haiku 5.5 small model and halves Sonnet 5.5 cache read prices. Checked
- Simon Willison: Claude Haiku 5.5. Checked
Alexandre S.
Alexandre S. is a creative ad strategist and copywriter with ten years' experience in digital marketing across multiple channels.