Anthropic shipped a new flagship model on Monday — and the sticker price didn’t move a cent.
Claude Sonnet 5.5, the second model in Anthropic’s Claude 5.5 family, went live on September 28, 2026. Pricing holds steady at $2 per million input tokens and $10 per million output tokens, unchanged from its predecessor, Sonnet 5. The pitch is efficiency instead: Anthropic says the new model generates output more than 30% faster and needs far fewer tokens to finish the same work, cutting the cost of a completed task by up to 30%.
“Introducing Claude Sonnet 5.5, the second model in the Claude 5.5 family,” Anthropic wrote on X. “It’s a clear upgrade over Sonnet 5, runs more than 30% faster, and costs up to 30% less for most work.”
That’s a rare pitch in an industry that usually sells progress as a higher bill.
The Claude Sonnet 5.5 price math: same tokens, fewer of them
Anthropic’s framing here is deliberate. The token price is unchanged: the savings come from the model doing more with less. For customers running coding agents or document pipelines, the cost of an individual job drops because Sonnet 5.5 burns fewer tokens to reach the same answer, and it arrives faster. TechCrunch noted that Anthropic’s own benchmarks show Sonnet 5.5 beating even Opus 5.5 on agentic coding, likely because it can spawn multiple agents without blowing the cost budget.
Early testers described the model as a better collaborator than Sonnet 5, according to Anthropic, with clearer writing and speed suited to quick iteration on less complex tasks. The company positions Sonnet 5.5 as its strongest model for everyday work with a clear scope: fixing bugs, producing polished documents, slides and spreadsheets, and refining user interfaces.
Cache reads stay at $0.20 per million tokens. The model is available on all major platforms (Amazon Web Services, Google Cloud and Microsoft Azure) under the model ID claude-sonnet-5-5, with zero data retention.
First Sonnet with Opus-grade cyber safeguards
The quiet headline in the launch may be the safety story. Sonnet 5.5 is the first Sonnet-class model to ship with cyber safeguards and fallbacks comparable to Anthropic’s most capable models. Its cyber capabilities are rated on par with Opus 5, the company said, while its biology safeguards remain unchanged from Sonnet 5, targeting a narrow set of high-risk requests and leaving routine software development and most life sciences work untouched.
That sequencing is unusual. Historically, the heavy guardrails arrived with the biggest models first. Bringing them down the stack to the mid-tier model suggests Anthropic expects Sonnet 5.5 to be used in places where the company wants those controls in place, including, per Reuters, an enterprise customer base that now accounts for about 80% of its business.
Those customers include Salesforce, Databricks, Goldman Sachs and Novo Nordisk. Enterprise has been the engine of Anthropic’s growth, and Sonnet — the workhorse tier between cheap-and-fast Haiku and heavyweight Opus, is the model most of those developers actually call.
Why this matters
The AI model race has a pricing problem. Token prices drift downward over time, but customers increasingly buy by the completed task, not the token: agents that run ten steps cost ten times whatever the token price says. Anthropic’s bet is that customers care more about the bill at the end of the month than the rate on the invoice.
There’s also a competitive wrinkle. OpenAI’s GPT-6 Sol launched on September 22 at the exact same $2 and $10 pricing, according to industry analysis. Google’s Gemini 3.8 Flash costs $0.75 and $3.75 through the end of the year, and open-weights models like DeepSeek’s V4.1 Flash run as low as $0.15 and $0.60 off peak. Sonnet 5.5 isn’t competing on price per token — it’s competing on getting the job done in fewer tokens, faster.
If that efficiency claim holds up in independent testing, it puts pressure on every rival mid-tier model. If it doesn’t, customers will notice quickly: task-level costs are much harder to hide than token prices.
The IPO backdrop
The timing is hard to miss. Anthropic is expanding its product lineup ahead of a planned IPO, Reuters reported. Sonnet 5.5 follows Claude Opus 5.5, which launched last week at $4 per million input tokens and $20 per million output tokens: a flagship for complex work that requires careful judgment. A third model, Haiku 5.5, built for fast, high-volume work, is due in the coming weeks.
It’s an odd week for Anthropic to be sprinting. CEO Dario Amodei earlier this month publicly called on the global AI community to slow down the pace of releasing new capabilities to address safety concerns. The company is now shipping its second frontier-class model in seven days. The tension between “preach restraint” and “ship fast” is the story of the AI industry in miniature.
What’s next
Haiku 5.5 lands soon, completing the 5.5 trio. Beyond that, the question is whether Anthropic’s efficiency pitch becomes the industry’s default frame, where everyone argues about cost per task instead of cost per token, or whether a rival undercuts it first.
For now, the math is simple enough for a finance team to understand: same price per token, 30% less work per dollar, 30% faster answers. In enterprise AI, that’s a harder pitch to beat than any benchmark chart.
Frequently asked questions
What is Claude Sonnet 5.5?
Claude Sonnet 5.5 is Anthropic’s new mid-tier AI model, released September 28, 2026. It’s a faster, more efficient successor to Sonnet 5 aimed at everyday coding and office tasks.
How much does Claude Sonnet 5.5 cost?
API pricing is unchanged from Sonnet 5: $2 per million input tokens and $10 per million output tokens. Anthropic says the model uses fewer tokens per task, so most work costs up to 30% less.
Is Claude Sonnet 5.5 faster than Sonnet 5?
Yes. Anthropic says Sonnet 5.5 generates output more than 30% faster than Sonnet 5, making it the fastest Sonnet model the company has shipped.
How does Claude Sonnet 5.5 compare to Opus 5.5?
Opus 5.5, launched the week before at $4 and $20 per million tokens, remains Anthropic’s flagship for complex work. Sonnet 5.5 is positioned as the faster, cheaper complement for well-scoped everyday tasks.
Sources: Reuters, TechCrunch, 9to5Mac