- Claude Sonnet 4.5 is now available
- Anthropic has announced the launch today in parallel with improvements in other existing products
- Claude Code now has control points and you can create files directly from Claude’s cat
Anthropic has published Claude Sonnet 4.5, the next generation of his incredibly popular AI model.
The company calls Sonnet 4.5 “The best coding model in the world” and says it is the “strongest model to build complex agents” and to “use computers”.
The new upgrade to Claude is launching alongside other upgrades to the most popular anthropic products, including the control points for the Claude code so that you can save your progress and return to the previous states, as well as the execution of the code and the creation of files for the calculation sheets, the slides and the documents of your conversation. These capacities will be available on all paid levels.
In addition to these great announcements, Anthropic has also confirmed that Claude for Google Chrome will be available for all those who had already joined the waiting list.
Anthropic says that Sonnet 4.5 is the “most aligned model of the company” and affirms that the new upgrade “will considerably improve the behavior of the model, reducing behaviors such as sycophance, deception, search for power and the tendency to encourage delirious thought.”
Claude Sonnet 4.5 is available today, and the prices remain the same per million tokens as the previous Sonnet 4 model.
A huge leap forward
Openai’s latest study shows that Claude Sonnet 4.1 BAT GPT-5, Gemini and Grok in real job tasks. Now, Anthropic says that Sonnet 4.5 shows “clear progress” compared to its predecessor, and not only the new model achieves higher performance in references, but it is also “better to meet the practical commercial needs than its predecessor, allowing our customers to do more difficult problems and to be more creative.”
Sonnet 4.5 can now help users create personalized agents using natural language, and Anthropic even claims that the model can operate independently for 30 hours, a massive improvement compared to the 7 -hour capacity of Claude Opus 4.
Companies are already impressed
Anthropic shared some quotes from its customers during the revelation of Sonnet 4.5, and the model was encountered by the first adopters with complimentary praise.
Danny Wu, AI products manager in Canva, said: “Claude Sonnet 4.5 provides impressive gains on our most complex and long -context tasks – engineers in our code base for features and product research.
While Sean Ward, CEO of Iigent AI, said that the new model, “resets our expectations – it manages more than 30 hours of autonomous coding, releasing our engineers to fight against months of complex architectural work in considerably less time while maintaining consistency through massive bases.”
Claude Sonnet 4.5 seems to be one of the largest versions of the AI of the year, and even if we do not try it ourselves, it certainly seems incredibly promising.
