Anthropic has launched a new frontier model called Claude Sonnet 4.5, which it claims offers state-of-the-art performance on coding benchmarks. The company says Claude Sonnet 4.5 is capable of building “production-ready” applications, rather than just prototypes, representing a leap in reliability from previous AI models. The new model is available via the Claude API and in the Claude chatbot. Pricing for developers is the same as the previous model, $3 per million input tokens and $15 per million output tokens. The company also launched a new Claude Agent SDK to help developers build their own agents, and a temporary research preview called “Imagine with Claude” for Max subscribers, which shows the model generating software on the fly.
Real-World Performance and Capabilities
In the last year, Anthropic’s AI models have become a favorite among developers and enterprises due to their strong performance on software engineering tasks, with companies like Apple and Meta reportedly using Claude internally. Anthropic claims that Claude Sonnet 4.5 offers industry-leading performance on coding benchmarks, including SWE-Bench Verified, a benchmark that measures real-world software coding abilities by presenting the model with failing tests and requiring it to propose a patch that actually fixes the bug.
Anthropic AI researcher David Hershey has said that during early trials with enterprise customers, he has seen Claude Sonnet 4.5 code autonomously for up to 30 hours, maintaining focus on complex, multi-step tasks. In one instance, he watched the AI model not only build an application but also set up database services, purchase a domain name, and perform a SOC 2 audit to ensure the product was secure. This is a significant leap in capability from the previous Claude Opus 4 model, which could run autonomously for only seven hours.
A Rapidly Evolving AI Landscape
The launch of Claude Sonnet 4.5 comes less than two months after Anthropic’s last AI model, Claude Opus 4.1. This rapid pace of development makes it difficult for any company to hold a meaningful lead for very long. This is especially true since OpenAI’s recent GPT-5 has challenged Anthropic’s dominance in the coding space, outperforming previous Claude models on a variety of benchmarks. In a statement, Cursor CEO Michael Truell said that Claude Sonnet 4.5 represents state-of-the-art coding performance, especially on longer tasks.
Anthropic is also claiming that Claude Sonnet 4.5 is its most aligned frontier AI model yet, with lower rates of sycophancy and deception. The company says it has also improved Claude’s susceptibility to prompt injection attacks, one of the most serious risks for users.
Author’s Opinion
This rapid pace of development and the impressive new capabilities of Claude Sonnet 4.5 signal a new, more complex phase in the AI race. The fact that a model can not only write code but also perform multi-step, real-world tasks autonomously for hours on end is a game-changer. It suggests that AI is moving beyond a simple writing assistant and into the role of a true co-pilot, capable of handling entire projects from start to finish. This new level of autonomy could fundamentally change how software is developed and could create a massive advantage for the companies that can master this technology, making the ongoing competition for enterprise clients more intense than ever before.
Featured image credit: Goran Ivos via Unsplash
For more stories like it, click the +Follow button at the top of this page to follow us.