Anthropic launches Claude Haiku 5.5 with 75% lower average costs

Anthropic launches Claude Haiku 5.5 with 75% lower average costs

The release includes cheaper Sonnet cache reads and monthly API credits for Max and Team subscribers.

Anthropic launched Claude Haiku 5.5, a small AI model designed for fast, high volume workloads that the company says costs roughly 75% less to run on average than Haiku 4.5.

The company describes Haiku 5.5 as its fastest model to date and its most capable small model. It targets tasks such as summarization, database queries, classification, live customer support and browser use.

For prompts of up to 100,000 tokens, Haiku 5.5 costs $0.10 per million input tokens and $0.50 per million output tokens. That compares with $1 and $5, respectively, for Haiku 4.5.

For prompts above that threshold, pricing rises to $0.50 per million input tokens and $2.50 per million output tokens. Anthropic said roughly 90% of requests to its previous Haiku model fell within the lower pricing tier.

Advertisement

Haiku 5.5 is also the first model in the Haiku family with an adjustable effort setting, allowing users to balance cost against reasoning capability.

Anthropic’s evaluations showed substantial improvements over Haiku 4.5. On the offline subset of OSWorld 2.1, which tests computer operation, Haiku 5.5 scored 72.4%, compared with 15.7% for its predecessor. On Terminal Bench 4.0, an agentic coding evaluation, it scored 39.2%, versus 70.6% for Sonnet 5.5.

The company said Sonnet 5.5 and Opus 5.5 remain better suited to complex coding tasks. Haiku is intended for narrower assignments and can work alongside the larger models as a coding subagent.

Alongside the launch, Anthropic halved Sonnet 5.5’s cache read price to $0.10 per million tokens. It estimates the change reduces costs by around 20% on most agentic workloads, where cached content accounts for a large share of token consumption.

Anthropic also announced monthly Claude Platform API credits for Max and Team subscribers, with rollout scheduled for this week. Max 5x users will receive $100 per month, while Max 20x users will receive $200. Team subscribers will receive up to $500 pooled across their users.

The credits can be used with any Anthropic model to build applications, tools and agents.

Haiku 5.5 is available through the Claude Platform, Amazon Web Services, Google Cloud and Microsoft Azure. Anthropic is also adding beta support for computer and browser use to its Python and TypeScript software development kits.

Disclosure: This article was edited by Estefano Gomez. For more information on how we create and review content, see our Editorial Policy.
Anthropic launches Claude Haiku 5.5 with 75% lower average costs
Anthropic launches Claude Haiku 5.5 with 75% lower average costs

The release includes cheaper Sonnet cache reads and monthly API credits for Max and Team subscribers.

Anthropic launched Claude Haiku 5.5, a small AI model designed for fast, high volume workloads that the company says costs roughly 75% less to run on average than Haiku 4.5.

The company describes Haiku 5.5 as its fastest model to date and its most capable small model. It targets tasks such as summarization, database queries, classification, live customer support and browser use.

For prompts of up to 100,000 tokens, Haiku 5.5 costs $0.10 per million input tokens and $0.50 per million output tokens. That compares with $1 and $5, respectively, for Haiku 4.5.

For prompts above that threshold, pricing rises to $0.50 per million input tokens and $2.50 per million output tokens. Anthropic said roughly 90% of requests to its previous Haiku model fell within the lower pricing tier.

Advertisement

Haiku 5.5 is also the first model in the Haiku family with an adjustable effort setting, allowing users to balance cost against reasoning capability.

Anthropic’s evaluations showed substantial improvements over Haiku 4.5. On the offline subset of OSWorld 2.1, which tests computer operation, Haiku 5.5 scored 72.4%, compared with 15.7% for its predecessor. On Terminal Bench 4.0, an agentic coding evaluation, it scored 39.2%, versus 70.6% for Sonnet 5.5.

The company said Sonnet 5.5 and Opus 5.5 remain better suited to complex coding tasks. Haiku is intended for narrower assignments and can work alongside the larger models as a coding subagent.

Alongside the launch, Anthropic halved Sonnet 5.5’s cache read price to $0.10 per million tokens. It estimates the change reduces costs by around 20% on most agentic workloads, where cached content accounts for a large share of token consumption.

Anthropic also announced monthly Claude Platform API credits for Max and Team subscribers, with rollout scheduled for this week. Max 5x users will receive $100 per month, while Max 20x users will receive $200. Team subscribers will receive up to $500 pooled across their users.

The credits can be used with any Anthropic model to build applications, tools and agents.

Haiku 5.5 is available through the Claude Platform, Amazon Web Services, Google Cloud and Microsoft Azure. Anthropic is also adding beta support for computer and browser use to its Python and TypeScript software development kits.

Disclosure: This article was edited by Estefano Gomez. For more information on how we create and review content, see our Editorial Policy.