- Anthropic has launched two new Claude AI models called Claude Ops 4 and Claude Swant 4.
- Claude OPS 4 claim that coding is the best AI in the world
- Claude Swant 4 is a small, smooth model with a large upgrade of Swant 3.7 version.
Anthropic has unveiled the latest generation of Claude 4 of its AI models. The company has been proud that the New Claude Ops 4 and Cloud Swant 4 are in the upper part of the game for Model AI assistants, with unprecedented coding skills and the ability to work freely.
Claude Swant is a small model, but it is still an important upgrade in power from previous Swant 3.7. Anthropic clams are much better in following Swant 4 instructions and coding. Even the Gut Hub has adopted a new co -coding agent to strengthen. It is likely to be more widely used because it is a default model for a free level for cloud chat boot.
Claude Ops 4 is the flagship model for the best coding AI according to anthropic and speculation. It can also handle sustainable, multi -hour tasks, and break it in thousands of steps to meet it. Opse 4 also includes tested “Extended Thinking” feature on the previous models. The expansionist thinking model allows to immediately respond and keep in the middle of search engines and other tools unless it has much data and can be resumed where it is left.
This means more than just long answers. Developer Ops 4 can train to use all kinds of third party tools. Opse 4 can also play video games very well, anthropic shows how AI performs when the file is allowed to access and make its navigation guide.
Claude 4 Power
Both Claude 4 model tools feature the use of the tool and enhance the concentrated properties around the memory. Ops 4 and Sonate 4 can use tools in the switches between parallel and logic and search. And when providing access to external files, their memory system can save and remove key facts over time. You will not need to re -explain what you want on every third signal.
To ensure that the AI ​​is doing your will, but in every detail does not overwhelm you, the model of Claude 4 also offers what is called “thinking”. Instead of the text wall of the text of thousands of steps, the Taken, to complete a prompt, has used a small, secondary AI model to reduce the train of thinking.
One of the supplementary benefits of working methods for new models is that they are less likely to be cheated to save time and processing power. Anthropic said he has reduced the behavior of shortcuting in the tasks that persuade AIS to make a fake way (or make something).
Big picture? Anthropic AI utility, especially coding and agent, is clearly counting for the lead in independent work. Chat GPT and Google Gemini have large user bases, but anthropic has at least some AI chat boot users to persuade users. Swant 4 free users and OPS 4 are available for bundling in Claude Pro, Max, team, and enterprise projects, trying to appeal to both fans of the Anthropic Budget friendly and Premium AI.