Anthropic has established itself as one of the main names in the AI industry. In recent weeks, companies like xAI and Perplexity have announced their advances in terms of better models or more powerful AI tools. Naturally, Anthropic couldn’t be left behind and has announced Claude 3.7 Sonnet as its first hybrid reasoning AI model.
Anthropic’s Claude 3.7 Sonnet is another intermediate step before AGIs
AI companies seem to be unmistakably heading towards developing products with greater “reasoning” capabilities. This gives them “agentic” qualities in preparation for a future where AGIs become a reality. Like other recent releases in the industry, the latest Claude model boasts greater complex problem-solving capabilities. It also outperforms its predecessors in key areas like math and coding.
Claude app users already have access to the updated 3.7 Sonnet model. It is also available to developers via the Anthropic API, Amazon Bedrock, and Google Cloud’s Vertex AI. There is no price increase, which is good news.
An all-in-one approach for better accessibility
Anthropic is taking an approach that AI companies aim to follow this year. Firms like OpenAI have released their own independent deep reasoning models. However, this can be confusing for those uninitiated in the space. Many won’t know which model to opt for based on their needs. To solve this, Anthropic wants to offer a one-in-all experience. This means that all of Claude’s capabilities will be available from a single model.
“We fundamentally believe that reasoning is a feature of the AI rather than a completely separate thing,” said Anthropic product research lead Dianne Penn. She also teases a fast response speed. Penn says the model shouldn’t take much longer to respond to the “What time is it?” prompt compared to the “Plan a two-week trip to Italy while considering the weather in late March.” prompt.
Updated knowledge database
Unlike ChatGPT, Gemini, and others, Claude can’t yet access the web. However, Anthropic has updated the AI platform’s knowledge base to a more current date. Plus, developers can set a response timeout in their apps built on Claude. “Sometimes the developer just needs to say it shouldn’t take more than 200 milliseconds to answer this question,” said Anthropic’s VP of product, Michael Gerstenhaber.
Claude Code, an improved coding tool
There’s also Claude Code, a new tool that, as its name suggests, focuses on coding. The company says it is “an active collaborator that can search and read code, edit files, write and run tests, commit and push code to GitHub, and use command line tools.” The tool will be available in a “limited research preview” form first.
Claude 3.7 Sonnet seems to have gotten ahead of everyone in terms of accessibility. Having all the possibilities of an AI platform in a single model will make things easier for the less tech-savvy. Recent leaks about OpenAI’s GPT-5 suggest that Sam Altman’s firm is going in the same direction. In this case, GPT-5 could integrate the o3-mini reasoning models. Grok 3 is another AI that makes its capabilities of both deep reasoning and deep research easy to reach. You just have to tap on one of the two icons before sending the prompt.
