Claude Opus 5 is now available, with Anthropic positioning the model as a major upgrade in intelligence and efficiency for software engineering, knowledge work, problem-solving and scientific research.

The company said Claude Opus 5 delivers performance approaching the frontier intelligence of Claude Fable 5 at half the price. The model is also the new default model for Claude Max subscribers and the strongest model available to Claude Pro users.

On coding and knowledge work evaluations, including Frontier-Bench and GDPval-AA, Claude Opus 5 reportedly achieves state-of-the-art performance. However, the model remains behind Mythos 5 on cybersecurity tasks.

The model is designed to improve performance while offering greater efficiency compared with its predecessor, Claude Opus 4.8. Users can adjust the model’s effort setting to balance intelligence, token usage, speed and cost.

On Frontier-Bench v0.1, Claude Opus 5 surpasses other models and more than doubles the performance of Opus 4.8 at a lower cost per task. On CursorBench 3.2, the model at maximum effort performs within 0.5% of Claude Fable 5’s peak score while costing approximately half as much per task.

The model also demonstrated strong performance on knowledge work and problem-solving benchmarks. On ARC-AGI 3, which tests a model’s ability to solve novel problems, Opus 5 achieved a score three times higher than the next-best model.

On Zapier AutomationBench, which evaluates the ability of AI models to complete business tasks from start to finish, Opus 5 recorded a pass rate approximately 1.5 times that of the next-best model at the same cost per task. Even at its lowest effort setting, the model reportedly completed more tasks than other models.

The model also outperformed other models across cost levels on OSWorld 2.0, a computer-use benchmark, surpassing the peak performance of Claude Fable 5 at just over one-third of the cost.

Claude Opus 5 also represents an improvement in scientific research capabilities, according to Anthropic. The model outperformed Opus 4.8 across the company’s life sciences evaluations, covering areas such as structural biology, organic chemistry and bioinformatics.

The strongest improvements were reported in organic chemistry and protein-related tasks. On an internal benchmark involving the inference of molecular structures from spectroscopy data, Opus 5 scored 10.2 percentage points higher than Opus 4.8. It also achieved a 7.7 percentage point improvement on tasks involving predictions of how variations in protein sequences affect their function.

With its combination of advanced reasoning, coding capabilities, scientific performance and adjustable effort settings, Claude Opus 5 is positioned as a model designed for everyday use as well as complex professional and research workloads.