Claude Haiku 5.5 Cuts AI Costs and Matches GPT-6 Luna

Anthropic has launched Claude Haiku 5.5 on its platform and major cloud services, positioning it as a fast, low-cost model for high-volume tasks such as summarisation, data lookup, classification and customer support. It offers five reasoning settings, from Low to Max, with Medium as the API default; Anthropic’s top benchmark results use Max, while complex agentic coding is better suited to larger models. For prompts up to 100,000 tokens, input costs $0.10 per million tokens and output costs $0.50, matching OpenAI’s GPT-6 Luna. Anthropic says this is one-tenth the price of Haiku 4.5 for prompts in that range, while its average operating-cost reduction is about 75%. Longer prompts receive a smaller discount, and a tokenizer change increases token counts by about 30%, so actual costs vary. Anthropic has also cut Sonnet 5.5 cache-read prices by 50%, estimating around 20% savings for many agent tasks. The launch sharpens competition in low-cost AI models, but it does not directly affect any cryptocurrency or token. Crypto traders may watch for broader effects on technology-sector sentiment and AI-related assets; any impact would likely depend on adoption and competitive developments rather than this announcement alone.