Anthropic launches Claude Haiku 5.5, its fastest and cheapest small model, and halves Sonnet 5.5 cache prices
Anthropic on October 7 released Claude Haiku 5.5, which it says costs around 75% less to run on average than Haiku 4.5 and is its fastest model to date, alongside a cache-price cut for Sonnet 5.5 and new monthly API credits for subscribers.

Anthropic on October 7 released Claude Haiku 5.5, which the company describes in its launch post as “the cheapest, fastest, and most capable small model we’ve ever released.”
According to Anthropic, Haiku 5.5 is designed for high-volume, cost-sensitive tasks such as summaries, compactions, database queries and classification requests. It is meant to pair with Opus 5.5 and Sonnet 5.5 as a subagent on coding work, and, as the company’s fastest model to date at standard speed, to suit speed-sensitive uses such as live customer support and browser use. Only Anthropic’s Opus models in Fast Mode run faster.
Pricing. Anthropic says Haiku 5.5 costs around 75% less to run than Haiku 4.5 on average. It is priced 90% lower for requests of up to 100,000 tokens, which made up about 90% of Haiku 4.5 traffic, and 50% lower for longer requests. Input costs $0.10 per million tokens ($0.50 above 100,000 tokens) and output $0.50 per million ($2.50 above 100,000). Haiku 4.5 had cost $1 and $5 respectively, The New Stack noted.
Haiku 5.5 is the first Haiku-class model with an adjustable effort setting, letting users trade off cost against intelligence. In early customer testing cited by Anthropic, HubSpot said the model posted the best score it had seen on its CRM evaluation suite, at 92.8%, while Box said it scored 11 points higher than Haiku 4.5 at about half the latency.
Safety. Anthropic says Haiku 5.5 shows major improvements over Haiku 4.5 across almost all of its alignment evaluations. Its cybersecurity safeguards are more restrictive than Haiku 4.5’s but somewhat less restrictive than those on other recent models: they permit a wider range of defensive tasks but still block penetration testing and other techniques more likely to be used by attackers.
Other changes. Anthropic also halved the price of cache reads on Claude Sonnet 5.5, from $0.20 to $0.10 per million tokens, which it says makes Sonnet 5.5 around 20% cheaper on most agentic work. This week it is rolling out monthly API credits for Max and Team subscribers on the Claude Platform: $100 a month for Max 5x users, $200 for Max 20x users and up to $500 for Team subscribers, pooled across users. Its Python and TypeScript SDKs are gaining beta support for computer use and browser use.
Haiku 5.5 is available now on the Claude Platform as claude-haiku-5-5, as well as on Amazon Web Services, Google Cloud and Microsoft Azure. AWS announced availability on Amazon Bedrock the same day, and The Verge highlighted the model’s speed. According to The New Stack, it is the first new version of Anthropic’s smallest model in nearly a year and the company’s third 5.5 model in a month.
Related articles

US backs $1.5B conditional loan for chips used in drones, radar, missile defense
The US Department of War has offered semiconductor manufacturer Wolfspeed up to $1.5 billion in...

World’s first fleet-capable drilling robot hits 10x speed with 99% accuracy
A drilling robot built for data center construction has already punched more than 385,000 holes...

US microreactor that reached criticality in 150 days set for full-power test in 2027
A US-developed nuclear battery that achieved its first controlled chain reaction in roughly 150 days...

Nous Research raises $90 million at $1.5 billion valuation to take its open-source Hermes agent to businesses
Nous Research, developer of the open-source Hermes Agent, has raised a $90 million Series B led by Robot Ventures and is launching Hermes for Businesses, letting companies deploy customized AI agents while keeping their data private.

Japanese wholesaler Nippan confirms book sales to Anthropic, deepening publisher concerns over AI training
Nippan Group Holdings has told the Japan Book Publishers Association that a subsidiary sold books to Anthropic from December 2024 to December 2025 but was not aware the purchases were for AI training, according to Jiji Press and Kyodo News.

AI mood sobers at Singapore forums as Temasek, Ray Dalio and officials flag market and existential risks
At the Milken Institute Asia Summit and the Forbes Global CEO Conference in Singapore on October 7, investors and policymakers broadened AI worries beyond valuations to loss of control, AI-designed bioweapons and whether huge spending will pay off.