Anthropic today launched Claude 3.5 Sonnet. The new AI model improves on its predecessor with faster performance and better skills in coding, vision, and natural language understanding.
While Sonnet will ultimately be Claude"s mid-tier model positioned between smaller Haiku and high-end Opus, Anthropic claims it outperforms even the top-of-the-line Opus based on internal benchmarks.
Claude 3.5 Sonnet processes inputs twice as fast as Opus, according to Anthropic. It says Sonnet outscored the previous Opus model on coding challenges, fixing 64% of code bugs compared to 38% for Opus.
Benchmark results also showed Sonnet beating competitors across seven of nine overall categories and four of five vision tasks. Though numbers should be taken cautiously, the scores indicate meaningful advances.
Claude 3.5 Sonnet is our strongest vision model yet, surpassing Claude 3 Opus on standard vision benchmarks. These step-change improvements are most noticeable for tasks that require visual reasoning, like interpreting charts and graphs. Claude 3.5 Sonnet can also accurately transcribe text from imperfect images—a core capability for retail, logistics, and financial services, where AI may glean more insights from an image, graphic or illustration than from text alone.
To ensure safety, Anthropic submitted Sonnet to external evaluations by the UK and US Artificial Intelligence Safety Institutes. They found the model remained at ASL Level 2 after enhancements. Anthropic also updated the assistant using child safety expertise to prevent potential harm better.
In addition to Claude 3.5 Sonnet, the company also introduced a new feature called Artifacts. With Artifacts, users can now see and interact with the results of their conversations with Claude directly within the app. For example, if a user asks Claude to design a flyer for an upcoming event, it can provide a preview of the flyer design for the user to review and edit.
This preview feature marks Claude’s evolution from a conversational AI to a collaborative work environment. It’s just the beginning of a broader vision for Claude.ai, which will soon expand to support team collaboration. In the near future, teams—and eventually entire organizations—will be able to securely centralize their knowledge, documents, and ongoing work in one shared space, with Claude serving as an on-demand teammate.
Like GPT-4o, Claude 3.5 Sonnet is free to use without limitation for casual, non-commercial purposes. However, heavy users will hit rate limits, as you discovered. For continued unrestricted access, subscription plans start at $20/month for individuals and scale up for businesses/enterprises.