Claude Haiku 5.5 launches at $0.10 per million tokens and is now Claude Code's Haiku

Anthropic released Claude Haiku 5.5 in the first week of October 2026, and Claude Code picked it up almost immediately. According to the Claude Code changelog, version 2.1.293 added claude-haiku-5-5 as the default Haiku model on the Anthropic API, with a 1M-token context window and pricing of $0.10 per million input tokens and $0.50 per million output tokens. Prompts over 100K tokens are priced at $0.50 and $2.50.

What is new

Press coverage, such as YourStory's write-up, describes Haiku 5.5 as Anthropic's cheapest and fastest small model and says it is the first Haiku with an adjustable effort setting. Reports put it roughly 75% cheaper than Haiku 4.5 on average. Outlets disagree slightly on the exact announcement date (October 7 or 8), so treat the day as approximate. The benchmark figures in that coverage come from secondary reporting rather than a page we could open on Anthropic's site, so we leave them out here.

The intended jobs, according to the coverage, are high-volume work: summaries, classification, support replies, browser tasks, and acting as a subagent beneath a larger model such as Sonnet 5.5 or Opus 5.5.

Why Claude Code users should care

The same week's Claude Code releases lean into that idea. The changelog (and the Releasebot digest) shows several subagent-related additions:

• Version 2.1.292 added an effort parameter to the Agent tool.
• Version 2.1.296 added autoCompactWindow to subagent frontmatter, so a subagent can compact its context earlier than the main conversation.
• Version 2.1.296 also added the CLAUDE_CODE_WORKFLOW_SUBAGENT_MODEL environment variable, which runs every workflow agent on one model while other subagents keep theirs.

Taken together, these changes make it easier to keep an expensive model for planning and hard edits while cheap models handle the repetitive background work.

What to do with it

If you already use subagents, check which model they run on. Search-and-summarize helpers are good candidates to move to Haiku 5.5. Keep code-writing and review agents on a stronger model until you have compared the output on your own repository. Pricing and model names change quickly, so confirm current numbers in your account before you budget around them.