Anthropic releases Claude Haiku 5.5, cutting small-model API prices by up to 90%
Input now costs $0.10 per million tokens for prompts up to 100,000 tokens. The benchmark gains over Haiku 4.5 are Anthropic’s own figures.
How this story was checked
- ResearchedWith AI assistance · 2 sources
- AI source checkEach sentence matched to its source
- Editor approvalMayank Sahu · 11 Oct, 20:44
- Published11 Oct, 20:44
In short
Anthropic released Claude Haiku 5.5 on 7 October 2026. For prompts up to 100,000 tokens it costs $0.10 per million input tokens and $0.50 per million output tokens, 90% less than Haiku 4.5. It is on the Claude API, AWS, Google Cloud and Microsoft Azure.
What launched
Anthropic released Claude Haiku 5.5 on 7 October 2026.1 It is available through the Claude API as claude-haiku-5-5 and, Anthropic says, on Amazon Web Services, Google Cloud and Microsoft Azure.12 Anthropic calls it its cheapest and fastest small model so far, and the first Haiku model with an adjustable effort setting.1
The new prices
For prompts up to 100,000 tokens, Anthropic lists:1
- Input: $0.10 per million tokens (Haiku 4.5: $1.00)
- Output: $0.50 per million tokens (Haiku 4.5: $5.00)
- Cache reads: $0.01 per million tokens (Haiku 4.5: $0.10)
Above 100,000 tokens the rates are $0.50 for input and $2.50 for output, a 50% cut.1 Anthropic says a new tokenizer uses slightly more tokens per task, so it estimates average costs about 75% lower than Haiku 4.5.1 VentureBeat notes that the lower rates match OpenAI’s published GPT-6 Luna prices.2
Anthropic also halved cache-read prices for Claude Sonnet 5.5, to $0.10 per million tokens.12
Benchmark claims
Anthropic reports large gains over Haiku 4.5. For example, it gives 72.4% on an offline subset of OSWorld 2.1, against 15.7% for Haiku 4.5.12 These are Anthropic’s own results and have not been checked independently. Anthropic says Sonnet 5.5 and Opus 5.5 remain the better choices for complex agentic coding.1
Safety settings
Anthropic says Haiku 5.5 has stricter cybersecurity safeguards than Haiku 4.5. They allow more defensive security work but still block penetration testing under standard settings.12 VentureBeat reports that wider access is available by applying to Anthropic’s verification programmes.2
Also announced
Max and Team subscribers get monthly API credits: $100 on Max 5x, $200 on Max 20x and up to $500 shared on Team.12 The Python and TypeScript SDKs add beta support for computer use and browser use.1
What it means for developers
Teams that run high-volume tasks such as sorting, extraction or support replies on Haiku 4.5 may see lower bills. Because of the new tokenizer, it is worth measuring token counts on real workloads before switching.
Disclosure: Anawords uses Claude, made by Anthropic, as one of its research tools. Anthropic has no say in our coverage.
Fact ledger
How labels workQuestions readers ask
How much does Claude Haiku 5.5 cost?
$0.10 per million input tokens and $0.50 per million output tokens for prompts up to 100,000 tokens; $0.50 and $2.50 above that.
Is Haiku 5.5 better than Haiku 4.5?
Anthropic reports large gains on its own benchmarks. Independent results were not available when we checked.
Where can I use it?
Through the Claude API as claude-haiku-5-5, and on AWS, Google Cloud and Microsoft Azure.