Anthropic Introduces Claude Haiku 5.5

Anthropic introduced Claude Haiku 5.5, a small model Anthropic describes as its cheapest and fastest yet, designed for high-volume, cost-sensitive tasks such as summaries, compactions, database queries, and classification requests.

For prompts up to 100,000 tokens, input tokens cost $0.10 per million and output tokens cost $0.50 per million, with cache reads priced at $0.01 and cache writes at $0.125. For prompts over 100,000 tokens, input tokens cost $0.50 per million and output tokens cost $2.50 per million, with cache reads at $0.05 and cache writes at $0.625. Alongside this launch, Anthropic halved the price of Claude Sonnet 5.5 cache reads to $0.10 per million tokens. The company also rolled out monthly API credits for Claude Max and Team subscribers, providing $100 per month for Max 5x users, $200 per month for Max 20x users, and up to $500 pooled for Team subscribers.

For computer use on OSWorld 2.1 (offline subset), it scored 72.4% versus 15.7% for Haiku 4.5 and 48.9% for GPT-6 Luna. On Humanity's Last Exam multidisciplinary reasoning, Haiku 5.5 scored 45.9% without tools and 57.4% with tools, compared to 10.2% and 18.7% for Haiku 4.5. On Terminal-Bench 4.0 agentic coding, it achieved 39.2% compared to 0.0% for Haiku 4.5 and 16.4% for GPT-6 Luna. On FrontierCode 1.1 (Main), it scored 46.4% against GPT-6 Luna's 42.4%. For visual reasoning on Chartography without tools, it scored 46.4% compared to 6.4% for Haiku 4.5 and 29.1% for GPT-6 Luna. Sonnet 5.5 reference scores on these same benchmarks include 83.9% on OSWorld 2.1, 56.9% without tools and 64.5% with tools on Humanity's Last Exam, 70.6% on Terminal-Bench 4.0, 52.1% on FrontierCode 1.1, and 61.6% on Chartography without tools.