Anthropic Releases 'Most Intelligent' LLM Yet: How Does It Compare To OpenAI, Deepseek?

The artificial intelligence industry is largely fragmented between many firms jockeying to be on the cutting edge of the technology. Progress is often benchmarked against competitors, as in the case of Anthropic with its new Claude 3.7 Sonnet.

A Step Forward: Anthropic, the AI startup backed by Jeff Bezos, announced the model on Monday on its website. The San Francisco-based company called it its “most intelligent” yet.

Anthropic says the model has improved capabilities for coding and front-end web development. It also allows for an “extended thinking mode” that performs better than its standard model but takes longer to return a response to a prompt.

Anthropic also noted a change in its priorities for the models’ strengths: “…in developing our reasoning models, we've optimized somewhat less for math and computer science competition problems, and instead shifted focus towards real-world tasks that better reflect how businesses actually use LLMs,” the announcement said.

How it Compares: When compared against each other, large-language models are often more proficient in some tasks and less proficient in others. OpenAI, led by Sam Altman, and DeepSeek, a Hangzhou, China-based company that sent shockwaves through financial markets with its highly efficient LLM, are among Anthropic’s competitors.

Anthropic benchmarked Claude 3.7 Sonnet against OpenAI and DeepSeek’s flagship models according to its own analyses.

The model with extended thinking beats OpenAI’s o1 and o3-mini and DeepSeek’s R1 in virtually every category besides math problem-solving. In particular, Claude 3.7 Sonnet is particularly strong in coding, easily beating other models. It is also strong in graduate-level reasoning and interactions with external systems.

Also Read:

• EXCLUSIVE: 3 ETFs Positioned For Microsoft’s Quantum Breakthrough

Image: Shutterstock