Claude Opus 5.5 launches: What’s new, how much it costs and what it can do
Claude Opus 5.5 launches
Anthropic has launched Claude Opus 5.5, a new artificial intelligence model aimed at delivering high-end performance while reducing the cost of running demanding AI workloads.
The model was unveiled on September 22, 2026, as competition among leading AI companies intensifies around coding, knowledge work, AI agents and enterprise applications. Anthropic says Opus 5.5 can deliver performance comparable to its higher-end Claude Fable 5.1 while costing about 40% less to operate on typical workloads.
That combination of capability and lower cost could make Opus 5.5 particularly relevant to businesses and developers that use AI extensively rather than only for occasional chatbot queries.
What is Claude Opus 5.5?
Claude Opus 5.5 is Anthropic’s latest Opus-class AI model, designed for complex software development and knowledge-intensive tasks.
Anthropic’s Opus models sit within the company’s Claude family and are aimed at demanding workloads where users need stronger reasoning, coding and agentic capabilities.
The new model follows Claude Opus 5, which Anthropic released in July 2026. The company described Opus 5 as a model designed for everyday use while offering strong performance in coding and professional work.
Opus 5.5 arrives as Anthropic continues expanding its Claude 5 family. The company’s Fable 5.1 and Mythos 5.1 models were introduced earlier in September, with different safety configurations for general use and specialised cybersecurity and life-sciences work.
Why is Opus 5.5 attracting attention?
The headline feature is not simply that Anthropic has released another model. It is the company’s emphasis on performance at a lower cost.
According to Anthropic, Opus 5.5 performs comparably to Fable 5.1 while costing about 40% less to run than its predecessor on typical workloads. Reuters also reported that Anthropic says the model outperformed OpenAI’s GPT-5.6 Sol on a software-development benchmark while costing roughly one-third as much to run.
Benchmark results should be read in context, however. Performance can vary considerably depending on the task, benchmark methodology, model settings and workload. Anthropic’s claims do not by themselves establish that Opus 5.5 is superior for every use case.
How much does Claude Opus 5.5 cost?
Anthropic has priced Opus 5.5 at $4 per million input tokens and $20 per million output tokens, according to Reuters.
That represents a reduction from the previous Opus pricing, with Reuters reporting that the new rates are about 20% below Claude Opus 5.
For developers running large numbers of AI requests, token costs can become a significant part of an application’s operating expenses. Lower pricing could consequently make it more practical to use a powerful model for larger workloads.
The actual cost for an individual user or company will depend on how much they use the model, the type of requests they send and the platform through which they access it.
What can Claude Opus 5.5 do?
Opus 5.5 is targeted at tasks requiring substantial reasoning and technical capability.
Potential uses include:
- Software development and code generation
- Debugging and reviewing code
- Complex knowledge-work tasks
- AI agents and automated workflows
- Research and analysis
- Enterprise applications
- Cybersecurity-related research under appropriate access programmes
- Life-sciences research
Anthropic has increasingly positioned Claude as a tool for professional and scientific workflows. Earlier in September, the company said Claude was responsible for 26% of its own AI research and development work, according to an analysis cited by Reuters.
Anthropic has also demonstrated Claude’s use in scientific work, including a project in which Claude helped formalise a computer-checked version of Fermat’s Last Theorem.
READ ALSO
Anthropic researchers warn AI could threaten human survival: Why safety measures may not be enough
Claude Fable 5.1 is more than a coding upgrade: Why Anthropic’s new AI model is drawing attention
Opus 5.5 comes with a major safety focus
The launch comes at an unusual moment for Anthropic because the company has simultaneously been raising concerns about the speed at which frontier AI capabilities are advancing.
Anthropic CEO Dario Amodei recently called for the AI industry to slow the release of new capabilities while researchers work on safety issues, according to Reuters.
Anthropic said Opus 5.5 underwent external testing by independent AI safety research groups Frontier Design and METR before its release. The company also said the model incorporates safeguards previously reserved for its most capable systems.
The company reported that Opus 5.5 performed better than previous systems in its internal safety evaluations and was about 85% less likely than Opus 5 or Mythos 5.1 to attempt to bypass containment boundaries in one dedicated evaluation. Those are Anthropic’s internal test results rather than a universal measure of AI safety.
Where can users access Claude Opus 5.5?
Anthropic says Opus 5.5 is available through major cloud platforms, including Amazon Web Services, Google Cloud and Microsoft Azure, according to Reuters.
Its availability through major cloud providers is significant for businesses that already build their AI infrastructure around those platforms.
Anthropic is also expanding programmes that provide vetted cybersecurity and life-sciences researchers with broader access to its models. The company separately launched a Life Sciences Verification Program in September to give approved organisations access to Claude models with safeguards adapted for biology-related work.
What happens to Claude Sonnet and Haiku?
Opus 5.5 is not the only model coming from Anthropic.
The company said Claude Sonnet 5.5 and Claude Haiku 5.5 will be released in the coming weeks, bringing similar performance, speed and safety improvements across other parts of the Claude model family.
That means Anthropic is building a broader 5.5 generation rather than treating Opus 5.5 as an isolated release.
For users, the distinction between the models will matter because different workloads can favour different combinations of intelligence, speed and cost.
Is Claude Opus 5.5 better than previous Claude models?
Anthropic says Opus 5.5 improves on previous systems in performance, cost efficiency and safety.
However, “better” depends on what a user needs. A developer running large volumes of relatively simple requests may value lower cost and speed, while someone handling complex research or software-engineering tasks may prioritise reasoning performance.
Anthropic’s own model lineup also shows why comparisons need context. Its Fable and Mythos models are designed with different safety configurations and access levels, while Opus models are positioned for highly capable general professional use.
Frequently asked questions about Claude Opus 5.5
What is Claude Opus 5.5?
Claude Opus 5.5 is Anthropic’s latest Opus AI model, designed for demanding coding, reasoning and professional knowledge-work tasks.
Is Claude Opus 5.5 free?
Access and pricing depend on the platform and plan being used. For API usage, Anthropic has announced pricing of $4 per million input tokens and $20 per million output tokens.
How much does Claude Opus 5.5 cost?
Anthropic’s announced API pricing is $4 per million input tokens and $20 per million output tokens, according to Reuters.
Is Claude Opus 5.5 better than GPT-5.6?
Anthropic reported that Opus 5.5 outperformed OpenAI’s GPT-5.6 Sol on a software-development benchmark. That result applies to the reported benchmark and should not be interpreted as proof that Opus 5.5 is better for every task.
What is Claude Opus 5.5 used for?
The model is intended for demanding tasks such as software development, coding, research, complex analysis, AI agents and enterprise workflows.
Where can I access Claude Opus 5.5?
Anthropic says Opus 5.5 is available through platforms including Amazon Web Services, Google Cloud and Microsoft Azure.
When were Claude Sonnet 5.5 and Haiku 5.5 released?
Anthropic said Sonnet 5.5 and Haiku 5.5 would arrive in the coming weeks following the Opus 5.5 launch.
Why is Anthropic focusing on cheaper AI models?
Lower operating costs matter because companies increasingly use AI at scale. Reducing the cost of powerful models can make larger deployments more economically practical, particularly for coding agents and enterprise applications.