Hermes Vector

AI & SI

Anthropic Releases Claude Haiku 5.5 and Lowers Sonnet 5.5 Pricing

Anthropic has released Claude Haiku 5.5, a small AI model priced around 75% less than Haiku 4.5 on average, and is halving cache read prices for Claude Sonnet 5.5.

By Hermes Vector

Also in: Español · Português · Français · Deutsch · 中文

Checked against 1 sourceAnthropic
Illustration for: Anthropic Releases Claude Haiku 5.5 and Lowers Sonnet 5.5 Pricing

In brief

  • Anthropic released Claude Haiku 5.5, described as its cheapest, fastest, and most capable small model.
  • Haiku 5.5 costs around 75% less to run on average than Haiku 4.5.
  • Cache read prices for Claude Sonnet 5.5 are halved to $0.10 per million tokens.
  • Claude Max, Ultra, and Team subscribers will receive monthly API credits of $100, $200, and up to $500 respectively.

Anthropic Releases Claude Haiku 5.5 and Lowers Sonnet 5.5 Pricing

Anthropic has released Claude Haiku 5.5, which the company describes as the cheapest, fastest, and most capable small model it has ever released. The model is designed for high-volume, cost-sensitive tasks such as summaries, database queries, and classification requests.

Performance and Pricing

On average, Claude Haiku 5.5 costs around 75% less to run than Haiku 4.5. The company states it is priced 90% lower than Haiku 4.5 for requests up to 100,000 tokens and 50% lower for requests over 100,000 tokens.

The price per 1 million tokens for Haiku 5.5 is:

  • Cache reads: $0.01 / $0.05
  • Cache writes: $0.125 / $0.625
  • Input tokens: $0.10 / $0.50
  • Output tokens: $0.50 / $2.50

The first price in each pair applies to prompts up to 100,000 tokens; the second applies to prompts over 100,000 tokens.

Alongside the launch, Anthropic is halving the price of Claude Sonnet 5.5’s cache reads. Cache reads now cost $0.10 per million tokens rather than $0.20, which the company says makes Sonnet 5.5 run around 20% cheaper on most agentic work.

Benchmarks

Anthropic provided benchmark results comparing Haiku 5.5 with Haiku 4.5, GPT-6 Luna, and Sonnet 5.5:

  • GDPval-AA v2.1: Haiku 5.5 scored 1620, Haiku 4.5 scored 735, GPT-6 Luna scored 1437, and Sonnet 5.5 scored 1840.
  • AA-Briefcase v1.1: Haiku 5.5 scored 1578, Haiku 4.5 scored 614, GPT-6 Luna scored 1336, and Sonnet 5.5 scored 1824.
  • OSWorld 2.1 (offline subset): Haiku 5.5 scored 72.4%, Haiku 4.5 scored 15.7%, GPT-6 Luna scored 48.9%, and Sonnet 5.5 scored 83.9%.
  • Humanity’s Last Exam: Haiku 5.5 scored 45.9% without tools and 57.4% with tools. Haiku 4.5 scored 10.2% without tools and 18.7% with tools. Sonnet 5.5 scored 56.9% without tools and 64.5% with tools.
  • Terminal-Bench 4.0: Haiku 5.5 scored 39.2%, Haiku 4.5 scored 0.0%, GPT-6 Luna scored 16.4%, and Sonnet 5.5 scored 70.6%.
  • FrontierCode 1.1 (Main): Haiku 5.5 scored 46.4%, GPT-6 Luna scored 42.4%, and Sonnet 5.5 scored 52.1%.
  • Chartography: Haiku 5.5 scored 46.4% without tools, Haiku 4.5 scored 6.4% without tools, GPT-6 Luna scored 29.1% without tools, and Sonnet 5.5 scored 61.6% without tools.

Availability and Credits

Claude Haiku 5.5 is available now on all platforms, including Amazon Web Services, Google Cloud, and Microsoft Azure. On the Claude Platform, developers can use claude-haiku-5-5.

Anthropic is also introducing a new monthly API credit for Claude Max, Ultra, and Team subscribers. Max users will get $100 in credits per month, Ultra users will get $200, and Team subscribers will receive up to $500, pooled across their users.

Why it matters

The release gives developers a lower-cost small model option for high-volume tasks, while the Sonnet 5.5 cache read price cut reduces costs for agentic work. The new monthly API credits are intended to support users building agents and applications on the Claude Platform.


Sources

This article was drafted with AI assistance and checked against the sources above. Company claims are reported as claims. Cover image is AI-generated.

Questions readers ask

What is Claude Haiku 5.5 and when was it released?
Anthropic has released Claude Haiku 5.5, which the company describes as the cheapest, fastest, and most capable small model it has ever released.
How much cheaper is Claude Haiku 5.5 compared to Haiku 4.5?
On average, Claude Haiku 5.5 costs around 75% less to run than Haiku 4.5, with pricing 90% lower for requests up to 100,000 tokens and 50% lower for requests over 100,000 tokens.
What changes were made to Claude Sonnet 5.5 pricing?
Anthropic is halving the price of Claude Sonnet 5.5’s cache reads to $0.10 per million tokens, which the company says makes Sonnet 5.5 run around 20% cheaper on most agentic work.
How do the new monthly API credits work for Claude subscribers?
Max users will get $100 in credits per month, Ultra users will get $200, and Team subscribers will receive up to $500, pooled across their users.

The day's checked news, by email

One short email a day with the top stories. Free, and you can unsubscribe in one click.

Or follow for the latest news:

How do you feel about this story?

Comments 0

No comments yet. Start the discussion.

AI & SI

Tech Executives Sign Petition Calling for Jensen Huang's Return to Nvidia

Elon Musk and other tech executives have signed a petition urging Nvidia to reinstate former CEO Jensen Huang, who resigned on November 12, 2026, following reports about his relationship with an alleged Chinese intelligence asset.

Checked against 5 sourcesTechCrunchindianexpress.comthefederal.com+2 more