Anthropic has launched Claude Opus 5.5 with lower operating costs, faster output and improved coding performance as competition among major AI developers intensifies. The company says the new model costs about 40% less to run than Claude Opus 5, with pricing of $4 per million input tokens and $20 per million output tokens. Cached reads now cost $0.20 per million tokens, while output generation is more than 30% faster, according to Anthropic.
Claude Opus 5.5 is available today. https://t.co/lSCmTYw9eJ
— Anthropic (@AnthropicAI) September 22, 2026
Claude Opus 5.5 is designed for coding, research and professional knowledge work. Anthropic says the model can deliver performance close to its previous generation while requiring less spending for common tasks.
The pricing changes come as AI companies face growing pressure to improve models without allowing inference costs to rise at the same pace. Anthropic is competing with companies including OpenAI, Alibaba, Moonshot AI and DeepSeek, all of which are developing increasingly capable models at different price points.
Source: anthropic.com
Anthropic also reported gains across several coding and professional benchmarks. On Terminal-Bench 4.0, Claude Opus 5.5 scored 66.4%, compared with 52.3% for Opus 5. On CursorBench 4.0, it recorded 57.8%, while its score on GDPval-AA was 1,846 Elo. The company uses these tests to measure software development and professional knowledge performance.
Anthropic is also showcasing the capabilities of the new approach in long-running software projects. One tester finished a migration involving 680,000 lines of code in less than a day, according to the business. In less than three hours, another tester audited and fixed a 200,000-line codebase using Opus 5.5.
Anthropic said the same amount of work took more than 20 hours with Claude Opus 5. The examples are intended to demonstrate how lower costs and faster output could matter for companies using AI for large software projects.
The debut coincides with a broader discussion over how soon frontier AI systems should be created and implemented. Elon Musk, OpenAI CEO Sam Altman, and Anthropic CEO Dario Amodei have all advocated for more stringent regulations and a slower rate of development.
At the same time, the commercial race continues. The source material says US President Donald Trump has opposed rules that could slow US AI development, while saying the Justice Department would monitor the sector and act where necessary.
Additionally, Anthropic is getting ready for a prospective public market debut. According to estimates, the corporation has been preparing for a possible November IPO, with a valuation close to $2 trillion. Anthropic has not made the final IPO terms or a completed valuation public, although it may try to raise up to $100 billion.
Lower model costs could be important as Anthropic expands its enterprise business. Claude Opus 5.5 is available through Anthropic’s Claude platform as well as Amazon Web Services, Google Cloud and Microsoft Azure. The company also plans to release Claude Sonnet 5.5 and Claude Haiku 5.5 in the coming weeks.
For businesses, the main attraction of the new model may not simply be higher benchmark scores. Lower inference costs and faster processing could make it more practical to use powerful AI models for large and repeated workloads.
Stay informed with the latest trends in Web3, blockchain innovation, and cybersecurity updates at 3verseTV
You need to login in order to Like
Leave a comment