he artificial intelligence startup has just released Claude 3.7 Sonnet, a new model designed to think about users' queries for as long as needed. Anthropic
considers the new AI model the first hybrid reasoning model in the artificial intelligence industry since it has the capability to provide real-time answers that at the same time full of valuable details.
When using the Anthropic AI model to generate responses, users have the feature to activate the thinking for the new Claude 3.7 Sonnet for a long or short period of time according to their needs.
What is Anthropic AI? As some of you wondered, Anthropic represents a research and AI safety company, founded in 2021, that operates for the purpose of creating trustworthy and controllable artificial intelligence systems. The company has released in the past various powerful AI models and now has revealed the Claude 3.7 Sonnet which is considered to have more advanced capabilities compared with the previous model, Claude 3.5 Sonnet.
So, the new AI Claude 3.7 Sonnet has a cost of $3 for a million input tokens, and that means that users can type in about 750,000 words. The cost for output tokens is $15 per million, and these two costs can be compared with the industry’s competitors which have much smaller costs for the same services. The o3-mini model from OpenAI has a cost of $1.10 per million of input tokens and $4.40 per 1 million of output tokens. On the other hand, the most recent and disputable model, DeepSeek R1 requires $0.55 for a million input tokens and $2.19 for a million output tokens.
Subscribe to our newsletter
It should be mentioned that even though these two models have a significantly smaller cost for the input and output tokens, they are only reasoning models compared with the new Claude 3.7 Sonnet which is a hybrid model.
The other industry’s reasoning models such as Gemini, Grok, and others need more time and use more computing power to generate detailed responses. These models are trained to break the prompt into smaller steps in order to enhance the accuracy of the final outcome. It’s true that these reasoning models are not able to think exactly like humans think, but the process behind them is similar to human reasoning.
“Similar to how humans don’t have two separate brains for questions that can be answered immediately versus those that require thought. We regard reasoning as simply one of the capabilities a frontier model should have, to be smoothly integrated with other capabilities, rather than something to be provided in a separate model.” the Anthropic company stated in a blog post.
Along with Claude 3.7 Sonnet, Anthropic AI also launched a new coding tool called Claude Code. Available for now only as a research preview, this new tool allows developers to operate specific tasks in Claude from their terminal. Claude Code is capable of analyzing different coding projects with just simple prompts from users.
According to a spokesperson, Claude Code will be available only for a limited number of users initially, and after some time, the company will probably release it for public use.
By
George King
•
February 25, 2025 12:10 PM
