Anthropic cuts Sonnet 5.5 cache read price from $0.20 to $0.10 and adds monthly API credits: $100 for Max 5x, $200 for Max 20x, up to $500 shared by Team users

Your API costs for certain Anthropic AI tasks are cut in half and you may receive monthly credits.

Anthropic has announced updates to its Claude AI models, including the release of Claude Haiku 5.5, which offers significant performance gains over its predecessor at a substantially lower cost. Prices for Haiku 5.5 are up to 90 percent cheaper for most common prompt lengths, making it highly competitive, especially for tasks like agentic coding where it reportedly outperforms OpenAI's GPT-6 Luna. Alongside this, Anthropic has halved the cache read price for its Sonnet 5.5 model and introduced monthly API credits for various subscription tiers, offering up to $500 for Team users. These changes aim to make AI services more accessible and cost-effective for developers and businesses utilizing Anthropic's API.

Anthropic's dual strategy of releasing a faster, cheaper Haiku 5.5 model and reducing Sonnet 5.5 cache read prices signals an intensified price war in the LLM market, directly challenging competitors like OpenAI. The introduction of API credits further sweetens the deal for existing users, potentially increasing platform stickiness and encouraging experimentation with more advanced features or higher usage volumes. This move indicates a broader trend towards tiered AI offerings designed to capture different market segments, from cost-conscious startups to enterprises needing reliable, scalable solutions.

The performance benchmarks for Haiku 5.5, particularly in agentic coding, suggest a maturing market where smaller, specialized models can rival or even surpass larger, more general-purpose models from competitors on specific tasks. This empowers developers to select more cost-effective solutions without sacrificing performance for many applications. The increased token usage with the new tokenizer, however, warrants careful monitoring by users to accurately assess real-world cost savings compared to headline price reductions.