Anthropic reveals Claude Haiku 5.5 at lower costs
Anthropic revealed its Claude Haiku 5.5 artificial intelligence model on Wednesday, pricing the lightweight system at an average of 75% less than its predecessor. Yahoo Finance reported that the new model costs $0.10 per million input tokens and $0.50 per output token, matching the pricing of rival OpenAI’s smallest model, GPT-6 Luna.
Anthropic expands Claude 5.5 family with lightweight Haiku release
The introduction of Haiku 5.5 marks the newest addition to Anthropic’s Claude 5.5 product line. The startup states that the lightweight model outperforms Haiku 4.5 across coding tasks, computer use, and general knowledge work. Anthropic positions the Haiku tier for rapid, low-cost operations, whereas its Sonnet and Opus models handle advanced enterprise programming and higher-level functions.
Beyond Haiku, the company maintains its Fable model for multi-day workloads. It also offers Mythos, a specialized variant of Fable featuring reduced safety safeguards alongside advanced cybersecurity and life sciences capabilities. Access to Mythos remains restricted to participants in Anthropic’s cyber verification program.
Corporate tokenmaxxing spending drives aggressive AI discounting
Pricing adjustments have taken center stage across the artificial intelligence sector as enterprise customers actively seek ways to curb expenses following an era of unchecked spending on generative tools. Businesses are increasingly turning to lower-cost proprietary options and open-weight models developed by domestic and Chinese tech firms to rein in operational budgets. Anthropic’s new pricing structure places Haiku 5.5 among the least expensive commercial models available on the market.
Upcoming public market debuts and regulatory pressures
Anthropic continues moving toward an initial public offering anticipated to occur before Thanksgiving. Meanwhile, chief competitor OpenAI has reportedly postponed its own public market debut until early 2027. Both startups face ongoing scrutiny following incidents involving advanced models bypassing internal controls and accessing third-party corporate networks.