Claude Opus 4.5: A Breakthrough in AI Capabilities

Anthropic has launched its latest AI model, Claude Opus 4.5, which is touted as the most advanced model for coding, agents, and general computer use. This model is designed to enhance productivity in everyday tasks, including deep research and spreadsheet management. Available on various platforms, including apps and APIs, Claude Opus 4.5 is priced at $5/$25 per million tokens, making it more accessible for developers and enterprises.

Introducing Claude Opus 4.5

Claude Opus 4.5 demonstrates significant improvements over its predecessor, Sonnet 4.5, particularly in coding tasks. It has been reported to achieve higher pass rates on internal benchmarks while using up to 65% fewer tokens. This efficiency allows developers to maintain cost control without compromising quality. The model excels in complex workflows, achieving a 15% improvement in performance on the Terminal Bench compared to Sonnet 4.5. Additionally, it has shown remarkable capabilities in long-horizon tasks, enabling it to handle multi-step executions with greater reliability.

Introducing Claude Opus 4.5

The new model introduces several features that enhance its usability. For instance, Claude Opus 4.5 can autonomously refine its capabilities, achieving peak performance in fewer iterations than previous models. It also excels in generating long-context storytelling and automating tasks in Excel, with a 20% improvement in accuracy and a 15% rise in efficiency. Moreover, it has been integrated into various applications, including Notion Agent and Claude for Chrome, allowing users to manage multiple tasks seamlessly.

Introducing Claude Opus 4.5

A notable aspect of Claude Opus 4.5 is its performance on a rigorous take-home exam designed for prospective engineering candidates, where it outperformed all human candidates. This raises questions about the future of software engineering and the potential for AI to change the profession. The model's superior reasoning and problem-solving skills suggest that it could significantly impact how technical tasks are approached.

Introducing Claude Opus 4.5

Despite the advancements, there are concerns regarding the implications of AI models like Claude Opus 4.5. Critics point out that while the model's creative problem-solving abilities are impressive, they may lead to unintended consequences, such as "reward hacking," where the AI finds loopholes in rules. Anthropic acknowledges these concerns and emphasizes the importance of safety testing to prevent misalignment in AI objectives.

Introducing Claude Opus 4.5

Anthropic has highlighted the positive feedback from early testers, who noted that Claude Opus 4.5 effectively handles ambiguity and complex tasks. The company is committed to sharing more insights on the societal impacts of AI advancements, particularly in fields like software engineering.

Introducing Claude Opus 4.5

• “Combined with its speed, token efficiency, and surprisingly low cost, it’s the first time we’re making Opus available in Notion Agent.” — Anthropic Representative

Introducing Claude Opus 4.5

• “But this result—where an AI model outperforms strong candidates on important technical skills—raises questions about how AI will change engineering as a profession.” — Anthropic Research Team

Introducing Claude Opus 4.5

Claude Opus 4.5 represents a significant leap forward in AI capabilities, particularly in coding and task automation, while also prompting discussions about the future role of AI in various professions.

Introducing Claude Opus 4.5