Anthropic introduced Claude 3.7 Sonnet – a new AI model that provides users with immediate responses as well as more thoughtful ones.
It is billed as “the first hybrid AI reasoning model” in the industry, allowing the user to select how long it “thinks.” before providing an answer.
Anthropic announced the rollout on Monday as part of its broader strategy for simplifying AI interactions.
Unlike other AI models that require the user to choose between multiple versions based upon cost and capabilities, Claude 3.7 Sonnet combines rapid response with deeper reasoning in a single AI model.
Subscribers only.
Anthropic makes Claude 3.7 Sonnet accessible to developers and users alike, but only those who subscribe to premium plans have access to reasoning abilities.
Users of the free tier will get a version that does not have reasoning, but which, according to the company, still performs better than its predecessor Claude 3.5 Sonnet.
Pricing reflects the model’s advanced features.
Claude 3.7 Sonnet has a price of $3 for every million input tokens, and $15 for each million output tokens. This is higher than OpenAI’s o3 mini and DeepSeek R1 which both are reasoning models that lack Claude 3.7 Sonnet hybrid functionality.
AI Reasoning Gains Popularity in Industry
Claude 3.7 Sonnet is a move towards reasoning-based AI. Many labs have adopted this method as the traditional improvement in AI performance has slowed.
Other models such as OpenAI’s o3 mini, Google’s Gemini 2.0 Flash Thinking and xAI’s Grok 3 Think also break problems down into smaller steps in order to improve accuracy.
Anthropic imagines an AI which can automatically determine the amount of time spent on reasoning. This eliminates the need for people to adjust their settings manually.
Diane Penn, the company’s head of product development and research, stressed that reason should not be treated separately but rather as an integrated function.
We believe that reasoning is an integral part of the frontier model, not a separate one. “Our belief is that humans have a single mind for quick reactions and deeper reflection. Anthropic said in a recent blog that this unified approach creates an easier experience for the users.
AI Race: Competitive Edge
Claude 3.7 Sonnet is launched as Anthropic aims to improve its competitive position on the AI landscape by taking on OpenAI ChatGPT, and Google Gemini.
According to reports, the startup has been in discussions to raise $2 billion by investors such as Lightspeed and Google. This would value the company at around $60 billion.
Amazon already has invested 8 billion dollars in the company.
Mike Krieger, Anthropic’s product chief and co-founder at Instagram, believes that the hybrid model is a good way to simplify AI chatbot interaction.
Krieger said that models all have their own personalities.
It’s too much to expect consumers to decide on the model and how long it should last. “We want consumers to concentrate on their needs, and not the technical details.”
Anthropic, according to Krieger, will add a new feature that allows users to “budget time” when reasoning. This ensures efficient answers tailored for different tasks.
The company also plans to release a tool for coding that uses AI agents.
Follow suit
Anthropic is known for its ability to innovate products ahead of the competition.
OpenAI quickly followed suit. The company had been the first in introducing a feature that was widely available for AI agents.
OpenAI’s CEO Sam Altman hinted that a similar method could be used to simplify AI interactions.
Altman said in a post on X from February that OpenAI aims at eliminating the need for users of OpenAI to select between models. He wants to get back to a magical unified intelligence.
Claude 3.7 Sonnet’s hybrid approach, which combines Claude 3.7 Sonnet’s AI functionality with a user-friendly interface, positions Anthropic to be a leader in the field of AI reasoning.
The post Anthropic Claude 3.7 Sonnet’s new release can ‘think as long as you want before answering’ may change as updates unfold.
This site is for entertainment only. Click here to read more
