![]()
Claude 3.7 Sonnet: A New AI Model with Flexible Thinking
Anthropic has introduced Claude 3.7 Sonnet, an advanced AI model that can "think" about questions for as long as needed. It is the first hybrid AI reasoning model, capable of delivering both quick responses and more detailed, thoughtful answers. Users have the option to decide how long the model should "reason" for, depending on the complexity of the question.
Availability and Pricing of Claude 3.7 Sonnet
Claude 3.7 Sonnet is available on all plans, but only premium users can access the reasoning features. Free users will get the standard version of the model, which still outperforms its predecessor, Claude 3.5 Sonnet. The model costs $3 per million input tokens and $15 per million output tokens, making it more expensive than some similar models like OpenAI’s o3-mini and DeepSeek’s R1.
![]()
Performance and Features of Claude 3.7 Sonnet
Claude 3.7 Sonnet has been designed to handle real-world tasks more effectively, such as coding and problem-solving. It shows significant improvements over earlier models in handling complex codebases and user interactions. Tests like SWE-Bench and TAU-Bench show that Claude 3.7 Sonnet is more accurate than its competitors, making it ideal for coding and planning tasks. It also has a unique feature that reveals its thinking process, allowing users to see how it arrives at answers, although some parts may be hidden for safety reasons.
Claude 3.7 Sonnet’s Reasoning Capabilities
Claude 3.7 Sonnet is a pioneer in integrating reasoning into a single model, unlike other models that use separate systems for different tasks. This allows users to enjoy a more seamless experience without having to choose between different models. Anthropic aims for Claude to eventually "decide" how long to think about a problem without users having to adjust settings manually.
The official link of this news is https://www.anthropic.com/news/claude-3-7-sonnet.