Anthropic launched Claude Haiku 5.5 on October 7, 2026 as its newest small model for high-volume, cost-sensitive work. Anthropic says it costs around 75% less to run on average than Haiku 4.5, while adding a 1 million token context window and an adjustable effort setting. The model is aimed at tasks such as summaries, classification, compaction, database queries, browser use and support workflows rather than replacing Opus or Sonnet on every complex job.
What Claude Haiku 5.5 is for
Anthropic positions Haiku 5.5 as the small, fast member of its Claude 5.5 family. The company says it is designed for quick and repetitive workloads and that it pairs well with Opus 5.5 and Sonnet 5.5 as a subagent for coding work.
This makes the model especially relevant to agent systems where a larger model handles the main task but needs many smaller calls for extraction, summarization, context compaction or tool preparation.
Claude Haiku 5.5 API pricing
| Prompt size | Input | Output |
|---|---|---|
| Up to 100K tokens | $0.10 per million | $0.50 per million |
| Above 100K tokens | $0.50 per million | $2.50 per million |
Anthropic's lower tier is the headline change. The company says Haiku 5.5 is around 75% cheaper to run on average than Haiku 4.5. Actual application cost still depends on token use, caching, retries, tool calls and how often the model needs to be rerun.
What the 1M context window means
Anthropic lists a 1 million token context window for Haiku 5.5. A large context can be useful for document-heavy workflows, but it does not mean an application should send every available document on every request.
Good systems still retrieve the relevant material, keep prompts focused and use caching where it makes sense. Context capacity is a limit, not a reason to increase every prompt.
Adjustable effort changes how teams can use a small model
Haiku 5.5 adds an effort setting. Anthropic says medium is the default. The idea is to spend more or less reasoning effort depending on the task instead of treating every request as equally difficult.
For a production workflow, this can be useful when most requests are easy but a smaller share need more reasoning. Teams can route harder cases to higher effort or a larger model instead of paying the higher cost for every request.
Where Haiku 5.5 fits in an agent stack
- Use a larger model to plan a complex task.
- Use Haiku 5.5 for high-volume extraction, summaries or small subagent jobs.
- Validate the result before it becomes an input to a high-impact action.
- Escalate difficult or ambiguous cases to a stronger model or human reviewer.
This approach can reduce cost, but only if the smaller model is reliable enough for the specific task. Token price alone is not a quality benchmark.
Availability and limitations
Anthropic says Haiku 5.5 is available through the Claude Platform, with availability also expanding across major cloud platforms. Google Cloud documentation lists the model as generally available from October 7, 2026. GitHub Copilot also lists Haiku 5.5 as generally available for supported paid plans.
Model availability, regional support and pricing can change by platform. Developers should check the provider's current pricing and model documentation before changing production routing.
What developers should check before migrating
- Measure error rates on your own workload, not only vendor benchmarks.
- Compare token counts because a different tokenizer can change real costs.
- Test long prompts separately from short prompts because the price tier changes above 100K tokens.
- Check caching, batch and cloud-provider pricing separately.
- Keep a fallback model for tasks where Haiku is not reliable enough.
Frequently asked questions
What is Claude Haiku 5.5?
It is Anthropic's small model for fast, high-volume and cost-sensitive workloads.
How much does Haiku 5.5 cost?
Anthropic lists $0.10 per million input tokens and $0.50 per million output tokens up to 100K tokens, then $0.50 and $2.50 above that threshold.
Does Haiku 5.5 have a 1M context window?
Yes. Anthropic lists a 1 million token context window.
Sources
- Anthropic: Claude Haiku 5.5
- Google Cloud: Claude Haiku 5.5 model availability
- GitHub Changelog: Claude Haiku 5.5 in GitHub Copilot