Claude Opus 5.5 is live — cheaper API rates, clearer writing, higher Pro and Team limits

By Jason Sibley

What happened

On 22 September 2026, Anthropic released Claude Opus 5.5, the first model in its Claude 5.5 family. Anthropic says it matches Claude Fable 5.1 on most work and costs about 40% less to run than Opus 5 on typical workloads. API pricing is $4 per million input tokens and $20 per million output tokens (down from $5 and $25 on Opus 5), with cache reads at $0.20 per million. Anthropic also says it is raising five-hour usage limits on Pro, Max, Team and seat-based Enterprise plans, and giving subscription users a rate-limit reset they can save and use when they choose. The model is available across Claude’s own apps and on AWS, Google Cloud and Microsoft Azure under the ID claude-opus-5-5. Source: Anthropic — Claude Opus 5.5.

What it means for a small business

If your team already drafts emails, proposals or research notes in Claude, the practical change is operational, not theoretical: more headroom on paid seats, and lower token bills if you call the API from Zapier, Make or a custom tool.

Writing quality is the SMB-facing bit. Anthropic says Opus 5.5 puts the main point first and uses less jargon than Opus 5. That matters when you are polishing a client email or a one-page proposal and you do not want another round of “make this sound human”.

Treat the cost claims as Anthropic’s own figures against Opus 5, not as a guarantee for your bill. Actual spend still depends on how long your prompts are and how often staff regenerate answers. Cyber and biology work sit behind stronger safeguards (similar to Fable 5.1), so if anyone on the team uses Claude for security testing or life-sciences tasks, expect more routing to older models unless you are in Anthropic’s verification programmes.

For UK and Irish firms, nothing here requires a new vendor. It is a model switch and a pricing change on a tool many already pay for.

What you can do this week

1. Subscription users: Open Claude settings, select Opus 5.5 where available, and run three live jobs you already do (a customer email rewrite, a proposal outline, a meeting-notes summary). Keep the old output beside the new one and note where you still have to edit.
2. API or automation users: Update the model ID to claude-opus-5-5 in one non-critical workflow first. Compare token use and wall-clock time on the same prompt before you roll it out to invoice chasers or CRM drafts.
3. Owner or ops lead: Write a two-line rule for the team: customer-facing numbers and commitments still need a human check. Confirm nobody is feeding client personal data into a personal Claude account outside your work plan.

How we can help

If you want a short session on which Claude plan fits a five-person team, and how to keep client data out of the wrong accounts, our SMB AI sessions cover exactly that: tool choice, prompts that survive a busy week, and a simple policy you can stick on the wall.