What's New in Claude · Released 22 September 2026
The top model got cheaper, and easier to read.
Claude Opus 5.5, released on 22 September 2026, is the first model in Anthropic’s 5.5 family. Anthropic says it performs at the level of Claude Fable 5.1 on most work and costs 40% less to run than Opus 5 on typical workloads. Its developer documentation now tells most teams to start with Opus 5.5.
For a company, two changes matter more than the benchmarks. Its writing is clearer and puts the important information first, so its work is easier to check. And Anthropic raised the five-hour usage limits on Pro, Max, Team and seat-based Enterprise plans, so fewer people stop halfway through a long task.
What changed
claude opus 5.5 release what's new
What Anthropic shipped.
Each figure is Anthropic’s, linked in the sources below.
- Price
- $4 per million input tokens and $20 per million output tokens, 20% below Opus 5’s $5 and $25. Cache reads are $0.20 per million, 60% below Opus 5. Anthropic estimates 40% lower cost than Opus 5 on typical workloads at default settings.
- Capacity
- A 1M token context window by default and up to 128k output tokens, with output more than 30% faster than Opus 5.
- Writing
- Anthropic says early testers found its writing clearer and easier to follow, with the most important information up front.
- Usage limits
- Higher five-hour limits on Pro, Max, Team and seat-based Enterprise plans, plus a rate limit reset subscribers can save and use when they choose.
- Safety
- Anthropic reports it is much less likely than recent models to take hard-to-reverse actions, and more resistant than Opus 5 to prompt injection.
What it means
how will claude opus 5.5 affect my business
What it means for your company.
The work where firms get the most from Claude is long, careful reading: a credit agreement, a quarter of portfolio reporting, a stack of diligence material. That is Opus work, and it just got cheaper and easier to review.
- Leadership
- The most capable everyday model now costs less than the one it replaces. If seat cost or API spend held back a wider rollout, the arithmetic has moved.
- Your teams
- Clearer writing matters more than it sounds. The slowest part of using AI on real work is checking the output, and a draft that leads with its conclusion and cites its source is faster to check.
- Risk and compliance
- Better resistance to prompt injection matters the moment Claude reads documents and email from outside the firm, which is most of the work worth doing. It lowers the risk; it does not remove the need for a person to review what goes out.
- IT and engineering
- Through the API, thinking cannot be disabled on Opus 5.5: requests that try return a 400 error, and depth is set with the effort parameter instead. Forced tool use also returns a 400. Read the migration guide before switching.
This week
What to do this week.
Three things, in this order.
- Make Opus 5.5 the default for judgment work
- Tell your teams which tasks get Opus: anything they would hand to a senior colleague to read and form a view on.
- Re-check who was hitting limits
- If people stopped using Claude because they ran out mid-task, tell them the limits went up. Some of them stopped trying and will not notice on their own.
- Test your API workloads
- Before moving anything from Opus 5 to Opus 5.5 in production, run it against your own examples with the migration guide open.
Sources
Where every figure on this page comes from.
Facts, prices and dates are Anthropic’s own, from the pages below. What it means and what to do are our view, from the firms we work with. Written by By The People Technologies and reviewed by Fatjon (Tony) Kalemaj, COO and Head of Applied AI, before it was published on 13 October 2026.
- Anthropic: Introducing Claude Opus 5.5
- www.anthropic.com
- Claude Platform release notes, 22 September 2026
- platform.claude.com
- Claude models overview and pricing
- platform.claude.com
Questions
The short answers.
The questions people ask about this release, answered in full.
How much does Claude Opus 5.5 cost?
$4 per million input tokens and $20 per million output tokens on the Claude API, with cache reads at $0.20 per million. Anthropic says that works out to about 40% less than Claude Opus 5 on typical workloads.
Is Claude Opus 5.5 better than Claude Fable 5.1?
Anthropic says Opus 5.5 performs at the level of Fable 5.1 on most work, at a lower price. Its documentation recommends Fable 5.1 for the most demanding reasoning and long-horizon agentic work, or when Opus 5.5 at higher effort still falls short on your own tests.
Did Claude usage limits go up with Opus 5.5?
Yes. With the launch on 22 September 2026, Anthropic increased five-hour usage limits on Pro, Max, Team and seat-based Enterprise plans, and gave subscribers a rate limit reset they can save and use later.
Want this working inside your firm?
We run Claude workshops on your own documents and stay on as your teams adopt each release. Write to us and say what you are trying to get done.
Subject line: Claude: Claude Opus 5.5, the new default for serious work