How to Use Claude Opus 5.5: What Changed, Limits and API Cost

For subscribers, the bigger news isn't on a benchmark chart. Anthropic raised the five-hour usage limits on Pro, Max, Team and seat-based Enterprise plans with this release, and gave subscribers a rate limit reset you can save and use when you choose.

Everything below comes from Anthropic's own Opus 5.5 announcement and its Sonnet 5.5 announcement, which I read in October 2026.

What's new in Opus 5.5 compared with Opus 5

Area	What Anthropic says changed
Capability	A major step up from Opus 5, at the level of Fable 5.1 on most work
Writing	Clearer and easier to follow, with the most important information up front
Speed	Generates output faster than Opus 5
Cost	Cheaper per token and uses fewer tokens per task
Safety	Best score to date on Anthropic's automated behavioural audit, and more resistant to prompt injection
Thinking	No longer available with thinking mode switched off

How to start using Claude Opus 5.5

1. Open Claude on the web, desktop or mobile and sign in.
2. Open the model menu in the chat box and choose Opus.
3. Set the effort level for the job. Lower effort answers faster and uses fewer tokens, higher effort reasons for longer.
4. If you hit your five-hour limit mid-project, use your saved rate limit reset rather than waiting.
5. Developers: switch your integration to claude-opus-5-5 and read Anthropic's migration guide first.

The usage limit change for subscribers

Anthropic didn't just make Opus 5.5 cheaper on the API. It increased five-hour usage limits on Pro, Max, Team and seat-based Enterprise plans, and added a rate limit reset that subscribers can save and spend whenever they choose.

In practice, that reset is best kept for a deadline day. If you're in the middle of a long report or a big spreadsheet job when you hit the limit, that's the moment to use it. Check Claude pricing if you're deciding which plan gives you enough Opus time.

What Claude Opus 5.5 costs on the API

Item	Opus 5.5, per million tokens
Input tokens	$4
Output tokens	$20
Cache reads	$0.20
Fast mode input	$8
Fast mode output	$40

Those are Anthropic's figures in US dollars, as of October 2026. Fast mode is available in Claude Code and on the Claude Platform and runs at up to 2.5x speed. Anthropic points out that cache reads make up most of the cost of agentic and coding work, which is why the lower cache price matters more than the headline rate.

The developer catch: thinking can't be switched off

This one will break some integrations. Anthropic says Opus 5.5 is no longer available with thinking mode switched off. It also launches with preserved thinking, a safeguard that stops API users editing Claude's earlier context to extract its reasoning, which applies to API accounts created on or after 31 August 2026.

If your app relied on thinking being off for speed, test it before you switch models. That's where Sonnet 5.5's new between_tools setting comes in, which Anthropic introduced for anyone running Sonnet with thinking off.

When Opus 5.5 is the right pick, and when it isn't

Anthropic describes Opus 5.5 as built for complex work that needs careful judgment, and says it remains clearly stronger than Sonnet 5.5 at complex, open-ended work. Sonnet 5.5 is the faster, cheaper choice for well-scoped everyday tasks.

• Use Opus 5.5 for: long research reports, tricky analysis, big code changes and anything where a wrong assumption is expensive.
• Use Sonnet 5.5 for: routine documents, slides, spreadsheets, quick fixes and fast back-and-forth.
• Use Haiku for: high-volume, low-stakes work. My guide to Claude Haiku 5.5 covers it.

Opus 5.5 for reports and knowledge work

If you don't write code, this is the part to look at. On GDPval-AA v2.1, an independent test of real-world work across 44 occupations run by Artificial Analysis, Anthropic reports Opus 5.5 scoring 1846 Elo, ahead of Fable 5.1 and Opus 5. In an internal test, Anthropic asked Opus 5.5 to write a report on a company's quarterly results using only what it could find on a copy of the web, with every figure and quote checked against sources.

Anthropic also says that on Zapier's AutomationBench, which tests real business workflows across connected apps, Opus 5.5 beat Opus 5 at every effort level. If your work is research, reporting or multi-app admin, that's the evidence that matters.

Cybersecurity and biology tasks get rerouted

Opus 5.5 is the first Opus model to launch with safeguards similar to Fable 5.1's on cybersecurity, biology and distillation. You can still find and fix bugs as part of normal software work, but Anthropic says most cybersecurity tasks are rerouted to Opus 4.8. For biology research held back by those safeguards, vetted organisations can apply to Anthropic's Life Sciences Verification Program.

Opus 5.5 is available on all platforms, including Amazon Web Services, Google Cloud and Microsoft Azure, and like earlier Opus models it's available with zero data retention.

My verdict

Opus 5.5 is the model to reach for when the work is hard and the answer matters. For most small businesses, the practical wins are clearer writing and more room under the usage limit. Keep Sonnet 5.5 for the everyday jobs, save your limit reset for deadline days, and if you build on the API, plan for thinking being permanently on before you migrate.

Frequently asked questions

When was Claude Opus 5.5 released?

Anthropic announced Claude Opus 5.5 on 22 September 2026 as the first model in its Claude 5.5 family.

Is Claude Opus 5.5 better than Opus 5?

Anthropic calls it a major step up from Opus 5, performing at the level of Claude Fable 5.1 on most work, writing more clearly and costing less to run.

How much does Claude Opus 5.5 cost on the API?

Anthropic lists $4 per million input tokens and $20 per million output tokens, with cache reads at $0.20 per million, in US dollars as of October 2026.

Can I turn off thinking on Claude Opus 5.5?

No. Anthropic says Opus 5.5 is no longer available with thinking mode switched off, so test any integration that relied on it before you switch.

Did Opus 5.5 change Claude usage limits?

Yes. 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 when they choose.