Claude Opus 5.5 Matches Its Flagship at 40 Percent Lower Cost

Anthropic released Claude Opus 5.5 on 22 September 2026, the first model in its new Claude 5.5 family. It matches the performance of Anthropic's previous flagship model on most work while costing 40 percent less to run, a rare combination of better and cheaper that matters for any business already paying for AI tools.

What did Anthropic announce with Claude Opus 5.5?

Opus 5.5 performs at the level of Claude Fable 5.1, Anthropic's most advanced model, on most everyday and coding tasks, but it is built to run far more efficiently. Input tokens now cost 4 US dollars per million and output tokens 20 US dollars per million, both about 20 percent less than the previous Opus model. Cached content, which makes up most of the cost of long coding and research sessions, dropped 60 percent to 0.20 US dollars per million tokens. Anthropic also says the model responds more than 30 percent faster and writes in plainer, more direct language, something early business users flagged as a real problem with the previous version.

Our customers use Box AI on enormous amounts of content, so speed and cost are a top priority. In our evaluations, Claude Opus 5.5 used a third of the tokens Opus 5 did, and its answers were 40 percent less verbose without losing accuracy.from Yashodha Bhavnani, VP of AI Products, Box (Anthropic, 22 September 2026)

What can Opus 5.5 actually do for a business?

The model is aimed squarely at practical work, not just benchmark scores. Anthropic reports that one early tester audited and fixed a 200,000 line codebase in under three hours, a task that took the previous model over 20 hours. Accounting firm Deloitte Consulting reported the model caught 72 percent of known bugs in code reviews even at its lowest effort setting, against 56 percent for the previous model at its highest setting, with far shorter output.

• Financial and research work: an investment firm's own evaluation found the model caught an error in its own test instructions that no earlier model had spotted.
• Long, unattended jobs: testers ran the model on multi hour engineering tasks across several company repositories overnight without supervision.
• Everyday writing and summaries: Anthropic redesigned how the model communicates after users said its predecessor was too wordy for quick business writing.

Why does the lower price matter for a small business?

Yes, mainly because it lowers the cost of AI tools many businesses already use, right as adoption is still catching up. HSBC's 2026 Voices of Treasury report found that 76 percent of Malaysian respondents expect AI to be extremely or very useful to their finance functions within three years, above the 72 percent Asia Pacific average. Yet 82 percent of the same respondents admitted their own department is not keeping pace with AI adoption happening elsewhere in their organisation, with integration into existing systems cited as the biggest barrier by 69 percent.

There is a real appetite among Malaysia's treasury community to explore what new payment instruments and AI enabled tools can deliver. At the same time, treasurers recognise that the challenge is often less about the technology itself and more about how it is integrated into existing, fragmented legacy and ERP systems.from Anand Mukati, Head of Global Payments Solutions, HSBC Malaysia (Business Today Malaysia, 17 September 2026)

Also today

• OpenAI released cheaper GPT-6 Sol and Luna models on 22 September 2026, cutting factual mistakes roughly in half while reaching its flagship model's reliability at lower cost (TechCrunch).
• OpenAI also announced it will let third party groups run safety assessments during training and evaluation of new models, earlier than the post development testing that has been standard practice (TechCrunch).

What to do this week

If your business already uses Claude or ChatGPT for daily work, check whether the newest model has quietly lowered your bill, most providers apply price cuts like this automatically. If you have been holding off on AI because it felt expensive or too complicated to fit into your existing systems, that excuse gets weaker every month this price war continues. In our Master Claude AI for Business Productivity program, we walk business owners through exactly which tools to use for which task, and how to fit them into daily work without a costly technical overhaul.