Performance Breakthrough
Claude Opus 5 delivers a massive leap over its predecessor Opus 4.8 while maintaining the same price point. On Frontier-Bench v0.1, Opus 5 more than doubles Opus 4.8’s performance. On CursorBench 3.2, at maximum effort, the model performs within 0.5% of Fable 5’s peak score but at half the cost per task.
The model excels across multiple evaluation categories:
- ARC-AGI 3: Opus 5’s score is three times higher than the next-best model on novel problem-solving tasks
- Zapier AutomationBench: Pass rate is approximately 1.5x the next-best model for the same cost per task
- OSWorld 2.0: Outperforms every other model at any given cost, surpassing Fable 5’s best result at just over a third of the cost
- Scientific research: Shows improved performance on every life sciences evaluation, with a 10.2-point gain on organic chemistry tasks and 7.7-point gain on protein-related tasks
Real-World Results
Early-access customers reported significant improvements across diverse use cases:
- Devin (AI coding agent): Opus 5 approaches Fable-level performance on FrontierCode 1.1 at half the cost, with particular strength on difficult debugging and root-cause analysis
- Cursor: Near Fable 5 intelligence at Opus speed and cost on CursorBench
- Zapier: Topped AutomationBench leaderboard without spending more tokens than prior Claude models, achieving 100% on a full churn-prevention sequence where previous models failed
- Box: Outperforms Opus 4.8 by 8% overall, with 11% improvement in data analysis and 17% in due diligence workflows
Alignment and Safety
Anthropic’s automated behavioral audit found Opus 5 to be their most aligned model to date. It adheres to Claude’s Constitution better than Opus 4.8, Sonnet 5, or Fable 5, exhibits the lowest rates of deceptive behavior, and is the least susceptible to being tricked into misuse.
On cybersecurity tasks, Opus 5’s classifiers are proportionally less restrictive than Fable 5’s, intervening around 85% less often. While Opus 5 approaches Mythos 5 at finding cybersecurity vulnerabilities, it remains substantially behind on developing exploits from those vulnerabilities.
Availability
Claude Opus 5 is available today on all platforms. It’s also offered in Fast mode, running approximately 2.5 times the default speed at twice the base price. Two beta features launched alongside:
- Mid-conversation tool changes: Developers can change which tools Claude uses without invalidating the prompt cache
- Automatic fallbacks on the API: Requests flagged by safety classifiers can automatically route to another model instead of being blocked