Anthropic Launches Claude Sonnet 5.5 As IPO Plans Advance

At a Glance

Anthropic has launched Claude Sonnet 5.5, the second model in its Claude 5.5 family.
The model retains Sonnet 5 pricing at $2 per million input tokens and $10 per million output tokens, while Anthropic says it can cost up to 30% less per task.
Sonnet 5.5 is designed for everyday enterprise workloads, including coding, document creation, presentations and spreadsheets.
The launch comes as Anthropic advances plans for a potential initial public offering and amid continuing debate over AI safety.

Anthropic has launched Claude Sonnet 5.5, expanding its Claude 5.5 product lineup as the artificial intelligence company prepares for a potential initial public offering.

Anthropic describes Sonnet 5.5 as a faster, lower-cost complement to its higher-end Claude Opus 5.5, with the company targeting routine enterprise and professional workloads.

Key Statements

“Sonnet 5.5 is a faster, lower-cost complement to Claude Opus 5.5.”
— Anthropic

“We must slow the pace at which we improve the capabilities of AI models.”
— Dario Amodei, Anthropic CEO

Sonnet 5.5 is priced at $2 per million input tokens and $10 per million output tokens, the same token rates Anthropic set for Sonnet 5. The company says the new model generally requires fewer tokens to complete the same work, resulting in savings of up to 30% per task. Anthropic also says Sonnet 5.5 generates outputs more than 30% faster than its predecessor.

The model is aimed at everyday professional tasks, including software development, document creation, presentations and spreadsheets. Anthropic positions it below Opus 5.5 for complex, open-ended work requiring sustained judgment, while emphasizing its speed and efficiency for more routine workloads.

Model	Input / 1M tokens	Output / 1M tokens	Primary positioning
Claude Sonnet 5.5	$2	$10	Everyday professional and enterprise tasks
Claude Opus 5.5	$4	$20	Complex, open-ended work
Claude Haiku 5.5	Not yet announced	Not yet announced	High-volume, cost-sensitive tasks

Anthropic launched Claude Sonnet 5.5 on September 28 as the second model in its new Claude 5.5 family, following the release of Claude Opus 5.5 on September 22.

Opus 5.5 is priced at $4 per million input tokens and $20 per million output tokens, compared with Sonnet 5.5 at $2 and $10 respectively. Anthropic says Opus is designed for more complex work, while Sonnet is intended for faster, well-defined tasks.

Anthropic says Sonnet 5.5 improves substantially on its predecessor in coding and knowledge-work evaluations. In the company's testing, it scored 70.6% on Terminal-Bench 4.0, compared with 10.3% for Sonnet 5. Anthropic cautions that benchmark results represent only one aspect of model performance.

The company has also emphasized safety measures, saying Sonnet 5.5 launches with cybersecurity safeguards similar to those used with its more capable models.

The launch comes at a notable moment for Anthropic. CEO Dario Amodei called earlier this month for AI companies to slow the pace at which model capabilities are developed, proposing independent safety evaluations, greater coordination among leading AI companies and international cooperation.

At the same time, Anthropic is preparing for a potential IPO. Reuters reported that the company confidentially filed for an IPO in June and remains on track for a potential public listing later this year.

Anthropic has said its enterprise business is a major part of its operations, with Sonnet 5.5 specifically targeting the kinds of coding and knowledge-work tasks common in business environments.

A third model, Claude Haiku 5.5, is expected to join the family in the coming weeks. Anthropic says Haiku is designed for high-volume and cost-sensitive applications.

FYI

Anthropic's latest release illustrates the tension facing major AI developers: companies are continuing to expand model capabilities and commercial offerings while their leaders simultaneously call for stronger safeguards and more deliberate development. Amodei's call for a slower pace came only weeks before Anthropic's latest model releases, underscoring that the company's approach is focused on pacing and safety measures rather than stopping AI development altogether.

The commercial context is also significant. Anthropic is pursuing rapid expansion while preparing for a potential public listing, and its confidential IPO filing has highlighted both the enormous infrastructure costs of frontier AI and the risks associated with increasingly capable systems. Reuters reported that Anthropic has commitments exceeding $518 billion for AI infrastructure over the next decade, illustrating the scale of computing investment required to compete in the sector.