RismadarVoice Reporters
September 22, 2026
Artificial intelligence company Anthropic has launched Claude Opus 5.5, its latest advanced AI model, introducing performance improvements, operating cost and safety as competition among leading artificial intelligence developers intensifies.
The company said the new model delivers performance comparable to some of its most capable systems while reducing the cost required to operate it.
According to Anthropic, Claude Opus 5.5 is priced at $4 per million input tokens and $20 per million output tokens, representing a 20% reduction compared with the previous Opus 5 model.
The company also said the new model could deliver comparable high-end performance at substantially lower operating costs than its predecessor, potentially making advanced AI capabilities more accessible to businesses, developers and organisations deploying artificial intelligence at scale.
Anthropic further reported that Claude Opus 5.5 performed strongly in software development evaluations, including outperforming OpenAI’s GPT-5.6 Sol on one benchmark cited by the company while costing significantly less to operate.
The benchmark result represents Anthropic’s own reported assessment and comes as major AI companies continue competing across areas including coding, reasoning, speed, safety and the cost of running increasingly sophisticated models.
Beyond performance, Anthropic placed significant emphasis on safety measures surrounding the new system.
The company said Claude Opus 5.5 underwent external testing involving independent artificial intelligence safety research organisations Frontier Design and METR before its public release.
Anthropic also said the model incorporates safeguards that had previously been reserved for its most capable AI systems.
In internal safety evaluations, the company reported that Opus 5.5 performed better than previous systems and was approximately 85 per cent less likely than Opus 5 or Mythos 5.1 to attempt to bypass containment boundaries during a specialised evaluation.
The launch comes amid wider international debate about the rapid development of increasingly powerful artificial intelligence systems and the potential risks associated with releasing new capabilities faster than appropriate safety mechanisms can be developed.
Anthropic Chief Executive Officer Dario Amodei had earlier in September called for greater caution across the global AI industry regarding the pace at which increasingly capable models are developed and released.
Despite those concerns, competition within the sector continues to accelerate as technology companies seek to develop models capable of performing complex professional tasks while reducing the enormous computing costs associated with advanced AI.
Claude Opus 5.5 is being made available through major cloud computing platforms, including Amazon Web Services, Google Cloud and Microsoft Azure, expanding its accessibility to businesses and developers building AI-powered products and services.
Anthropic also announced that Claude Sonnet 5.5 and Claude Haiku 5.5 are expected to be released in the coming weeks.
The upcoming models are expected to incorporate several of the performance, speed and safety improvements introduced with Opus 5.5.
The company is also expanding programmes that provide vetted cybersecurity and life-sciences researchers with broader access to its artificial intelligence systems for specialised research.
The latest release adds another major model to the rapidly evolving global AI industry, where developers are increasingly competing not only on raw model capability but also on affordability, efficiency, safety and the ability to deploy advanced systems at scale.
