Claude 3.7 Sonnet Takes Back the AI Crown

Anthropic has unveiled its latest AI model. Unlike the approach taken by OpenAI and others that use specialized versions offering either traditional LLM or reasoning capabilities, Claude 3.7 Sonnet packs both under one roof in its “hybrid reasoning model.”

This means Claude 3.7 Sonnet can produce normal chat responses or extended, step-by-step thinking. Anthropic says it wants to simplify the experience of using a model, and designed it such that prompting works similarly in both modes.

Standard and extended modes

In standard mode, Claude 3.7 Sonnet is an upgraded version of Claude 3.5 Sonnet. In extended thinking mode, it self-reflects before answering. This improves its performancein math, physics, instruction-following, coding, and many other tasks. But while OpenAI offers its reasoning model to free users, Claude 3.7 reasoning capabilities, or “extended” mode, is only available to paying subscribers.

Based on results across various benchmarks, Claude 3.7 Sonnet is back in the lead in the white-hot race for AI supremacy, edging ahead in areas such as graduate-level reasoning, visual reasoning, and instruction-following. It also outperforms earlier Anthropic models in areas like math and coding.

This upgrade came together with a “limited research preview” of Claude Code, which is billed as “an active collaborator that can search and read code, edit files, write and run tests, commit and push code to GitHub, and use command line tools.”

“Claude 3.7 Sonnet and Claude Code mark an important step towards AI systems that can truly augment human capabilities. With their ability to reason deeply, work autonomously, and collaborate effectively, they bring us closer to a future where AI enriches and expands what humans can achieve,” wrote Anthropic in its blog.

Despite the improvements in the AI model, there are concerns that Anthropic is lagging in terms of features. Specifically, it still cannot generate images, is unable to browse the web, and lacks the deep research capabilities that OpenAI, Grok, and even Perplexity now offer.

Finally, it is worth noting that Claude 3.7 Sonnet is priced at a premium compared to its competitors. Specifically, API access is pegged at USD3 per million input tokens and USD15 per million output tokens. This is significantly higher than what OpenAI, Microsoft, and Google offer.

Image credit: Anthropic