Anthropic has introduced Claude Opus 5, a new AI model that combines near-flagship performance with improved cost efficiency.
The model is designed to meet growing enterprise demand for powerful yet affordable AI tools across coding, business productivity, and other knowledge-intensive tasks.
The San Francisco-based lab said that the new Claude AI model was well suited for daily office and computer programming tasks.
According to Anthropic, Claude Opus 5 has set a new state-of-the-art benchmark on coding and knowledge work evaluations, including Frontier-Bench and GDPval-AA, demonstrating its strong performance in software development and complex business tasks. However, the company noted that the model continues to trail its flagship Mythos 5 on cybersecurity-related benchmarks.
Opus 5 is designed for everyday enterprise use, delivering greater efficiency than many competing models while maintaining high performance. The company has made Opus 5 the default model for Claude Max subscribers and its most powerful offering available to Claude Pro users.
Claude Opus 5 delivers significantly improved performance at the same price as its predecessor, Claude Opus 4.8, allowing customers to balance intelligence, speed, and cost by adjusting the model’s effort settings to suit different workloads.
The company noted that Opus 5 performs particularly well on software engineering tasks. On the Frontier-Bench v0.1 benchmark, the model outperformed all competing AI models and achieved more than double the performance of Opus 4.8 while reducing the cost per task.
On CursorBench 3.2, Anthropic said Opus 5, when operating at its highest effort setting, scored within 0.5% of its flagship Fable 5 model while costing about half as much per task, offering higher performance across high, extra-high, and maximum effort settings than competing models at comparable cost levels, as per the statement.
