Claude Haiku 5.5: Price, Benchmarks, API ID and Limits

Quick answer: Anthropic released Claude Haiku 5.5 on October 7, 2026 as its cheapest and fastest small Claude model. For prompts up to 100,000 tokens, the API costs $0.10 per million input tokens and $0.50 per million output tokens. Above that threshold, input rises to $0.50 and output to $2.50 per million tokens. The model ID is claude-haiku-5-5, and Anthropic says it is available through the Claude Platform, AWS, Google Cloud and Microsoft Azure. Track this release alongside other major model launches in the Avarixo AI Model Tracker.

Claude Haiku 5.5 pricing at a glance

The biggest change is price. Anthropic positions Haiku 5.5 for high-volume work such as summarization, classification, database queries, compaction, customer support and narrow subagent tasks. The company says the model costs about 75% less to run on average than Haiku 4.5 once request mix and tokenizer differences are considered.

Anthropic says roughly 90% of requests to the previous Haiku model fell into the cheaper prompt-size tier. That makes the 100K boundary important when estimating production costs.

How much faster and stronger is Haiku 5.5?

Anthropic’s evaluation table shows a large generational jump. On the OSWorld 2.1 offline subset for computer use, Haiku 5.5 scored 72.4%, compared with 15.7% for Haiku 4.5. On Terminal-Bench 4.0, Haiku 5.5 scored 39.2% while Haiku 4.5 scored 0.0% in the published table. On GDPval-AA v2.1 knowledge work, the new model scored 1,620 versus 735 for its predecessor.

Those are vendor-reported evaluations rather than independent guarantees of performance on every workload. The practical takeaway is that Haiku is no longer positioned only as a basic routing or classification model; Anthropic is also pitching it for browser use, live support and small agent tasks where latency matters.

Haiku 5.5 vs GPT-6 Luna and Sonnet 5.5

At the lower pricing tier, Haiku 5.5 matches the published base input/output rates of GPT-6 Luna at $0.10 and $0.50 per million tokens. Anthropic’s table reports Haiku 5.5 ahead of Luna on the selected OSWorld and Terminal-Bench evaluations, while Sonnet 5.5 remains stronger on the more demanding agentic benchmarks shown by Anthropic.

That makes the models useful for different jobs. Haiku 5.5 is aimed at short, repeatable tasks that need to run many times. Sonnet and Opus remain more appropriate when the task is complex enough that reliability or deep coding ability matters more than per-token cost.

New effort control and developer features

Haiku 5.5 is the first Haiku-class model with an adjustable effort setting. Developers can trade some cost and speed for more careful reasoning instead of treating every request the same way. Anthropic is also updating its Python and TypeScript SDKs with beta support for computer use and browser use.

Where Claude Haiku 5.5 is available

The model is available on Anthropic’s platform and through Amazon Web Services, Google Cloud and Microsoft Azure. Developers using the native Claude Platform can call it with the model identifier claude-haiku-5-5.

Anthropic also tightened some cybersecurity safeguards compared with Haiku 4.5. The company says defensive work is supported more broadly than on some larger Claude models, while penetration testing and other higher-risk offensive techniques remain restricted outside verified programs.

Other Claude pricing changes

Sonnet 5.5 cache reads fall from $0.20 to $0.10 per million tokens. Anthropic estimates that this can reduce the cost of many agentic workloads by around 20% when prompt caching is heavily used.

Max and Team subscribers are also set to receive monthly API credits: $100 for Max 5x, $200 for Max 20x and up to $500 pooled across Team users.

Who should consider Haiku 5.5?

The strongest fit is a workflow that runs the same narrow task thousands or millions of times: document summaries, field extraction, classification, quick database lookups, support responses, routing, compaction and subagent work.

For more model changes, see AVARIXO’s AI & Tech coverage.

FAQ

What is the Claude Haiku 5.5 model ID?

claude-haiku-5-5.

How much does Claude Haiku 5.5 cost?

For prompts up to 100,000 tokens, published rates are $0.10 per million input tokens and $0.50 per million output tokens. Higher rates apply above that threshold.

Is Haiku 5.5 available on AWS and Google Cloud?

Yes. Anthropic says it is available through AWS, Google Cloud and Microsoft Azure as well as its own platform.

Sources

Anthropic — Claude Haiku 5.5