Anthropic Cut Claude Haiku's Price by 90% to $0.10 per Million Input Tokens, the Same List Price as OpenAI's GPT-6 Luna — and an Independent Test Found It Costs About Three Times as Much per Task

On Oct 7, 2026, Anthropic launched Claude Haiku 5.5 with two rate cards chosen by prompt length: $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens, and $0.50 and $2.50 for longer ones. Claude Haiku 4.5 costs $1 and $5 at any length, so the short-prompt rates are 90% lower; Anthropic says prompts of that size made up around 90% of requests to Haiku 4.5. The short-prompt rates are the same as OpenAI's GPT-6 Luna, launched Sept 22. Anthropic's own estimate is that Haiku 5.5 costs around 75% less to run than Haiku 4.5 on average, a figure that accounts for its newer tokenizer, which its migration guide says turns the same input text into approximately 30% more tokens. The independent benchmarker Artificial Analysis ran Haiku 5.5 and GPT-6 Luna, both at max effort, on its Intelligence Index: Haiku 5.5 scored 43 to Luna's 38 but used about 162,000 output tokens per task to Luna's roughly 50,000, for a cost of about $0.21 per task against $0.07, about three times as much. The catch is the house thesis in one launch: the price per token fell 90%, the cost per finished task depends on how many tokens the model spends. Anthropic has confidentially filed for an IPO.