In an internal agentic coding evaluation, Anthropic's new Claude 3.5 Sonnet solved 64% of problems, nearly doubling the performance of its predecessor, Claude 3 Opus. Claude 3.5 Sonnet's ability to solve 64% of problems, nearly doubling its predecessor's performance, allows developers to address significantly more intricate coding challenges, streamlining complex software development.

AI models rapidly increase in capability and speed. Claude 3.5 Sonnet, however, demonstrates these advancements do not require a trade-off in safety. The model operates at twice the speed of Claude 3 Opus, further enhancing its utility for urgent development tasks.

Claude 3.5 Sonnet's release will likely drive market demand for AI models offering both high performance and verifiable safety. Claude 3.5 Sonnet's release, by driving market demand for AI models offering both high performance and verifiable safety, pushes competitors to innovate on both fronts simultaneously.

A New Benchmark for Agentic Coding

  • In an internal agentic coding evaluation, Claude 3.5 Sonnet solved 64% of problems, while Claude 3 Opus solved 38%, according to AI Magazine.
  • Claude 3.5 Sonnet operates at twice the speed of Claude 3 Opus, according to AI Magazine.

This combination of nearly doubled coding problem-solving success and twice the operational speed allows developers to tackle significantly more complex and time-consuming tasks. It fundamentally shifts the economic calculus of AI-assisted development. Companies assuming advanced AI coding assistance is too risky or slow are now demonstrably behind; Claude 3.5 Sonnet's ability to solve 64% of coding problems at twice the speed of its predecessor, while maintaining an ASL-2 safety rating, makes it a non-negotiable tool for competitive software development.

Unwavering Commitment to AI Safety

Claude 3.5 Sonnet maintains an ASL-2 safety rating, according to red teaming assessments by AI Magazine. Claude 3.5 Sonnet's sustained ASL-2 safety level, alongside a dramatic leap in coding performance, challenges the industry assumption that increased AI capability necessitates a safety trade-off. Anthropic demonstrates that advanced performance and speed can be achieved without compromising established safety standards.