Quick answer
Anthropic released Claude Haiku 5.5 on October 7, 2026 for high-volume, cost-sensitive tasks. Its listed API rates are $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens, with higher rates above that threshold. Anthropic reports lower average cost than Haiku 4.5, but the new tokenizer uses about 30% more tokens for the same text and several API settings changed. These are provider-published details, not an independent model test; availability in Chat AI is unverified.
Haiku 5.5 targets short, frequent, and latency-sensitive work
Anthropic positions Haiku 5.5 for repeated tasks such as summaries, compaction, database queries, and classification, as well as speed-sensitive customer support and browser use. The company also describes it as a subagent option alongside Sonnet 5.5 and Opus 5.5 for coding workflows. Those are Anthropic's intended use cases, not a guarantee that Haiku will outperform a larger model on every task.
Sources: Anthropic
API rates are lower for shorter prompts, but token counts changed
Anthropic lists Haiku 5.5 at $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens. Above that threshold, the listed rates are $0.50 and $2.50. Anthropic says the model costs around 75% less to run on average than Haiku 4.5, but that estimate reflects the company's request mix. Its new tokenizer produces about 30% more tokens for the same text than Haiku 4.5, so teams should recount representative prompts and recalculate costs rather than comparing rates alone.
Sources: Anthropic, Claude Platform Docs
Check API settings before switching from Haiku 4.5
The migration guide lists several compatibility changes. Haiku 5.5 uses adaptive thinking instead of the older enabled-thinking budget, rejects assistant-prefill continuations, and no longer accepts custom temperature, top_p, or top_k values. The guide also calls out changes to computer-use tools and notes that large images can consume more visual tokens. A safe migration is to change the model ID in a test environment, recount prompts, update affected settings, and run the same representative tasks before moving production traffic.
Sources: Claude Platform Docs
Anthropic lists API and cloud access; Chat AI access is unverified
Anthropic says Haiku 5.5 is available through the Claude Platform and named cloud providers; its Claude API model ID is claude-haiku-5-5. Chat AI's published model directory does not contain an exact Haiku 5.5 entry, so this article does not claim that the model is selectable in Chat AI. Check the current in-app model list before relying on access. For a broader selection checklist, compare apps by the tasks, exact models, and limits you need rather than by provider names alone.
Frequently asked questions
What readers usually ask
What is Claude Haiku 5.5's API model ID?
Anthropic lists claude-haiku-5-5 for the Claude API. Check the current provider documentation for platform-specific identifiers before deploying.
How much does Claude Haiku 5.5 cost?
Anthropic lists $0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens. The rates rise to $0.50 and $2.50 above that threshold. Actual cost also depends on token counts and usage.
Does Haiku 5.5 always cost 75% less than Haiku 4.5?
No. Around 75% lower average cost is Anthropic's estimate based on its request mix, not a guaranteed saving for every prompt. The tokenizer also uses about 30% more tokens for the same text, so recalculate with representative inputs.
Is Claude Haiku 5.5 available in Chat AI?
Its availability is unverified. The published Chat AI model directory does not show an exact Haiku 5.5 entry; check the current in-app model selector rather than assuming access.
Evidence