Claude Haiku 5.5 Guide

Claude Haiku 5.5, released on October 7, 2026, is Anthropic’s fastest and cheapest model. It sits alongside Fable 5.1, Opus 5.5, and Sonnet 5.5 in the current Claude lineup.

It is designed for:

• High-volume tasks
• Fast responses
• Subagents that work under larger models
• Classification and extraction
• Summaries and routing
• Low-cost browser and computer-use agents

Pick the smallest model that does the job well. These days, that is often Haiku.

Context window

1 million tokens

Maximum output

128,000 tokens

Effort levels

Low through max, with medium as the default

Thinking

The first Haiku model with adaptive thinking

API model ID

claude-haiku-5-5

Prompts of up to 100K tokens

$0.10 per million input tokens and $0.50 per million output tokens

Prompts above 100K tokens

$0.50 input and $2.50 output, per million tokens

Cache reads

Starting at $0.01 per million tokens

Batch API

50% off

Haiku 5.5's base prices are about 90% lower than Haiku 4.5's ($0.10 vs $1 per million input tokens). One review found its tokenizer turns the same text into more tokens, so test your real workload.

Anthropic’s launch benchmarks show a large leap over Haiku 4.5:

OSWorld 2.1

72.4%, up from 15.7% on Haiku 4.5

Terminal-Bench 4.0

39.2%, up from 0.0%

Humanity’s Last Exam, with tools

57.4%, up from 18.7%

Sonnet 5.5 still beats Haiku 5.5 on every benchmark Anthropic published. These are vendor results, so test on your own tasks.

• Claude apps, which offer Haiku models on every plan
• Claude API
• Amazon Bedrock
• Google Cloud
• Microsoft Foundry

Anthropic has committed to keep Haiku 5.5 available until at least October 7, 2027.

• Tagging and routing support tickets
• Pulling fields from invoices and forms
• Summarizing long documents
• Working through moderation queues
• Powering search and retrieval helpers
• Running simple browser automation
• Producing fast drafts and rewrites

Anthropic positions Haiku 5.5 as a worker that reports to Sonnet 5.5 or Opus 5.5:

• The larger model plans and makes decisions
• Haiku reads files and collects information
• Haiku runs narrow subtasks in parallel
• The larger model reviews and merges the results

Big model for judgment. Haiku for volume.

Sonnet 5.5

Coding agents and complex, multi-step work

Opus 5.5

Hard reasoning and daily flagship work

Fable 5.1

Long-running agents and the toughest problems

• Switch the model ID to claude-haiku-5-5
• Swap manual thinking budgets for the effort setting
• Drop non-default temperature, top_p, and top_k settings
• Drop assistant prefill
• Update computer-use tool versions
• Rerun your evaluations before moving production traffic

Anthropic lists these as breaking changes. Read its migration guide.

• Keep prompts below 100K tokens where you can
• Cache repeated instructions and documents
• Send non-urgent jobs through the Batch API
• Choose low or medium effort for simple tasks
• Request short, structured outputs

• Weaker than Sonnet 5.5 as a coding agent
• Long prompts cost five times as much per token
• Safety classifiers may refuse some requests
• Still capable of mistakes, so check important outputs

You are a fast assistant for [task]. For each item, return only: [fields] in [format]. If information is missing or unclear, return “unknown” rather than guessing. Keep every answer under [length].