Anthropic
Anthropic’s hybrid AI is distinguished by the fact that it combines the capabilities of standard and reasoning models — that is, it can provide both quick answers and detailed answers with step-by-step reasoning visible to the user (interestingly, the API even allows you to specify how many tokens the model can spend on reasoning).
According to Anthropic (via TechCrunch), Claude 3.7 Sonnet demonstrates significantly noticeable improvements in coding and front-end development, as well as math. The company’s employees have successfully used the new model to create front-end website designs and interactive games
It is noted that the new model was also tested on the classic Game Boy game Pokémon Red — пrgarding to the previous version, Claude 3.0 Sonnet, which failed to leave the Pallet Town house where the story begins, Claude 3.7 Sonnet successfully fought with the three leaders and took their badges.
Claude 3.7 Sonnet is available in all pricing plans (including free, but in this case, the advanced reasoning mode will not work) and via the Anthropic, Amazon Bedrock, and Google Cloud Vertex AI APIs. Prices remain unchanged: $3 per million incoming tokens and $15 per million outgoing tokens, including «reasoning» tokens.
In addition to the new model, Anthropic has also released in «limited preview» its coding agent called Claude Code, which is positioned as an «active collaborator that can read code, edit files, commit and submit code to GitHub, and use» command line tools.
In the demo, Anthropic showed how Claude Code can analyze a coding project with a simple command like «Explain the structure of this project» and make edits with simple queries in English (the model will describe everything it does during the changes).
Claude Code will initially be available to a limited number of users on a first-come, first-served basis.