Claude 4.5 is available everywhere today. Through the API, the model supports the same price as the Claude Sonnet 4, in the amount of $ 3 per million input tokens and $ 15 per million tokens. Developers can get access to it through the API Claude using Claude-Sonnet-4-5 as a model identifier.
Other new functions
Some of the auxiliary features of the Claude family also received some updates today. For example, Anthropno added code execution and creating files directly in conversations for users of the Claude web interface and selected applications. In this direction, users can now generate electric tables, slides and documents without leaving the chat interface.
The company also released a five -day preliminary viewing of a study called “Imagine with Claude“For MAX subscribers that demonstrate a model that generates real -time software. Anthropic describes it as “a funny demonstration showing what Claud Sonnet 4.5 can do in combination with the corresponding infrastructure.
The ecrinth of available models of anthropic artificial intelligence for users Claude Max, which can be seen in the Claude web interface on September 29, 2025.
Credit: Benge Edwards
As mentioned above, the Claude Code command line developing the command line has also received several updates today, along with the new model. The company added checkpoints that retain progress and allow users to return to previous conditions, updated the terminal interface and sent their own code. The Claude API also receives a new function of editing context and a memory tool to perform longer tasks of the agent.
Currently, AI companies are especially clinging to control indicators of software development as evidence of the possibility of AI's assistant, because progress in other areas is difficult to objectively measure, and this is an area where LLMS may have shown high usefulness compared to other areas that may suffer from concern. But people are still using chat bots of artificial intelligence, as a Claude as common assistants. And given the recent news about problems with some users Having passed fantastic rabbit holes It is possible with the Chat bots of AI more than usual that the anthropic statements of the Claude Sonnet 4.5 demonstrate “Sicophalism, Deception, the search for power and the tendency to encourage delusional thinking” compared to previous models. SikophaneIn particular, it is a tendency for AI model to praise the user's ideas, even if they are wrong or potentially dangerous.
We could add to how the anthropic caps of some of this weekend behavior of AI through a resolutely anthropomorphic lens. we have in the pastBut in general, attempts to reduce Sikophany are welcome news in the world, which is increasingly addressing the chat bots for much greater than just helping in coding.
