Antarubor Claude Sonit 4.5, the best model for Amnesty International Code

On Monday, Antarbur launched a new border model Claude Sonit 4.5 is calledWhich claims to perform on the latest model on coding standards. The company says that Claude Sonnet 4.5 is able to build “ready -to -produce” applications, instead of the initial models, which represent a jump in reliability of previous artificial intelligence models.

Claude Sonit 4.5 will be available via API Claude and Claude Chatbot. The pricing of the developers is the same as Claude Sonit 4: $ 3 per million input symbols (approximately 750,000 words, or more than the entire Lord of the Rings) and $ 15 per million directing symbols.

Last year, artificial intelligence models appeared Favorite between developers and institutions, In a large part of it because of their strong performance in software engineering tasks. According to what was used Apple and Meta, Claude AI models have made Anthropor an important work to sell API to artificial intelligence coding applications such as the indicator, Windsurf and re -re -coding applications. Recently, Openai’s GPT-5 The anthropologist has challenged space, outperformed Claude models over a variety of coding standards.

Anthropier says The bench has been checked. However, David Horshey Techcrunch researcher David Horshey Techcrunch tells that it is difficult to pick up Claude Sonit 4.5 performance on standards alone.

Hershey says he saw the Claude Sonit 4.5 code independently for up to 30 hours during early experiments with some institutions agents. At that time, see the artificial intelligence model not only create an application, but stand up database services, purchase names, and the performance of SOC 2 to ensure that the product is safe.

In a joint statement with the CEO of Techcrunch Cursor Micheal Truell In the Claude Sonnet 4.5 represents the performance of modern coding, specifically in the long horizon tasks. “Claude Sonnet 4.5 is a” new generation of coding models. “

An anthropier also claims that Claude Sonit 4.5 is the most compatible artificial intelligence model so far, with a decrease in SYCOPHANCY rates and deception of previous models. The company says it has also improved Claude’s susceptibility to injection attacks.

Besides the launch of Clauds Sonnet 4.5, Antarbur also launches Clauds Agent SDK. The company says this is the same infrastructure that operates the Claude icon, and can be used to help developers to build their own agents.

Antarbur also launches a temporary research inspection called “Imagine with Claud” for MAX subscribers, which shows the artificial intelligence generation program during flying. The company says the model will respond to user requests in the actual time, with no pre -defined functions or a cubic code.

The strained competition in the world of artificial intelligence has made companies to charge pioneering models every few months. Claude Sonit 4.5 launch less than two months after the latest artificial intelligence model in Antarbur, Claude Obus 4.1. These are rapid production courses that make it difficult for any company to have meaningful progress for a very long time.