Anthropic has announced the release of Claude Opus 5, available today. This new model is positioned as a thoughtful and proactive AI, approaching the intelligence of Claude Fable 5 at half the price. It is designed for daily use and is now the default model on Claude Max and the strongest model on Claude Pro.
Claude Opus 5 provides enhanced performance for the same cost as Opus 4.8. The model's effort setting allows customers to optimize for intelligence or conserve tokens for faster and more cost-effective results.
Key Points
- Claude Opus 5 is available today and is the new default model on Claude Max and the strongest on Claude Pro.
- It approaches the intelligence of Claude Fable 5 at half the cost.
- On Frontier-Bench v0.1, Opus 5 surpasses other models and more than doubles Opus 4.8's performance at a lower cost per task.
- On CursorBench 3.2, at max effort, Opus 5 performs within 0.5% of Fable 5's peak score but at half the cost per task.
- Opus 5 shows improved performance on life sciences evaluations, including structural biology, organic chemistry, and bioinformatics.
- It scores 10.2 percentage points higher than Opus 4.8 on internal benchmarks for inferring molecular structures from spectroscopy data.
- Opus 5 scores 7.7 percentage points higher on protein-related tasks, such as predicting how variations in a protein's sequence affect its function.
Context
According to Anthropic, Claude Opus 5 represents a significant improvement for long-running agents and delivers advancements in coding and professional work. It achieves state-of-the-art results on coding and knowledge work evaluations like Frontier-Bench and GDPval-AA, though it remains behind Mythos 5 on cybersecurity tasks. The model is also noted for its ability to verify its work and iterate carefully, as observed in evaluations and early-access testing.
Why It Matters
Builders and researchers can leverage Claude Opus 5 for improved performance in coding, knowledge work, and scientific research tasks, potentially reducing operational costs while maintaining or enhancing output quality. The model's ability to approach higher-tier intelligence at a lower price point offers a new option for optimizing resource allocation in AI-powered workflows.
What To Do
- Compare the performance charts for Opus 5 against Opus 4.8 to understand the impact of the effort setting on intelligence and token usage.
- Test Opus 5 on coding tasks, particularly those involving debugging and root-cause analysis, to evaluate its performance against Fable 5-level capabilities.
- Note the specific improvements in scientific research, especially in organic chemistry and protein-related tasks, for relevant applications.
- Watch for Opus 5's behavior in agentic workflows, observing its ability to verify work and iterate carefully.