Now you can now deal with the Claud AI model

Photo of author

By [email protected]


Antarubor is working to increase the amount of information that corporate customers can send to Claude in one claim, which is part of an attempt to attract more developers to the famous artificial intelligence coding models of the company.

For customers of the application programming interface for anthropology, the company Claude Sony 4 The artificial intelligence model now contains a symbolic context window of one million million – which means that AI can deal with requests as long as 750,000 words, more than the “Lord of the Rings”, or 75,000 line of software instructions. This is the maximum five times before Claude (200,000 icons), and more than twice the window of the distinctive symbol 400000 provided by GPT-5 Openai.

A long context will also be available for Clauds Sonnet 4 through Cloud’s Cloud partners, including Amazon Bedrock and Google Cloud’s Vertex Ai.

The Anthrop built one of The largest companies for institutions Among the developers of artificial intelligence models, to a large extent by selling Claude to artificial intelligence coding platforms such as Microsoft Github Copilot, Windsurf and Anysphere’s Cursor. Although Claude has become a preferred model among developers, GPT-5 may threaten the dominance of Antarbur with its competitive pricing and the performance of strong coding. The CEO of Anysphere Michael Truell helped Openai GPT-5 launchIt is now the AI virtual model for new users in the index.

In an interview, Brad Abrams, who is provided by Anthropor products, told Techcrunch in an interview that he is expected to get artificial intelligence coding platforms on “a lot of benefits” from this update. When asked if the GPT-5 has placed a dent in the use of Claud’s API, ABRAMS reduced anxiety, saying that he was “really happy with API and the way you grow.”

While Openai generates most of its revenues from consumer subscriptions to ChatGPT, the Anthropic commercial centers on selling artificial intelligence models to institutions through the application programming interface. This has made artificial intelligence coding platforms as a major customer of Arthur, and the reason may be that the company is aimed at some new privileges to attract users in the face of GPT-5.

Last week, Anthropor revealed an updated version of the largest artificial intelligence model, Claude Obus 4.1, Which pushed the capabilities of the company’s artificial intelligence further.

TECHRUNCH event

San Francisco
|
27-29 October, 2025

In general, artificial intelligence models tend to perform better in all tasks when they have more context, but in particular for software engineering problems. For example, if you ask the artificial intelligence model to rotate a new feature of your application, it is likely to do better if he can see the entire project, instead of just a small section.

ABRAMS also told Techcrunch that the large context of Claude’s context helps it to perform better in the long coding tasks of the agent, as the artificial intelligence model works independently on a problem for minutes or hours. With a large context window, Claude can remember all his previous steps in long horizon tasks.

But some companies have taken large context windows to the extreme, as they claim that artificial intelligence models can handle huge claims. Google offers a symbolic context window of two million people to Gueini 2.5 Pro, and Meta offers a 10 million symbolic context window for Llama 4 Scout.

Some studies indicate an end to the effectiveness of the windows of the great context; Artificial intelligence models Not In addressing these huge claims. Abrams said that the anthropologist’s research team focused on increasing the window of context only Claude, but also the “effective context window”, indicating that artificial intelligence can understand most of the information it provides. However, he refused to detect the accurate anthropologist.

When the Claud Sonnet 4 claim exceeds more than 200,000, the Anthrop will collect more for users of the API, at a price of $ 6 per million input codes and $ 22.50 per million output symbols (from 3 dollars per million input symbols and 15 dollars per million distinct symbols).


We are always looking forward to evolution, and by providing an insight into your view and comments in Techcrunch, coverage and events, you can help us! Fill this poll to inform us how to do AGet a chance to win a prize in return!



https://techcrunch.com/wp-content/uploads/2023/09/53202070940_ea57312b1a_k.jpg?resize=1200,800

Source link

Leave a Comment