Anthropic Launches Upgraded Claude 3.5 Sonnet, New Haiku Model, and Experimental Computer Use
Anthropic has unveiled several key advancements to its Claude AI platform, including an upgraded Claude 3.5 Sonnet model, the introduction of Claude 3.5 Haiku, and a new computer use feature in public beta. These innovations mark significant strides in AI’s coding, automation, and interaction capabilities.
Claude 3.5 Sonnet: Improved Performance for Coding and Automation
The upgraded Claude 3.5 Sonnet delivers advanced agentic coding and improved tool use, excelling in complex tasks for development and automation. Early adopters, like GitLab and Cognition, have noted its enhanced reasoning and problem-solving power for DevSecOps and autonomous AI evaluations, outpacing all previous models.
Claude 3.5 Haiku: High-Speed, Cost-Effective AI
Anthropic also introduced Claude 3.5 Haiku, an upgraded, budget-friendly model matching the speed and affordability of Claude 3 Haiku. Claude 3.5 Haiku achieves exceptional results on coding benchmarks, showing strong performance in user-facing tasks and data-driven personalization.
Public Beta of Computer Use Capability
A groundbreaking computer use feature is now available in beta, allowing Claude to interact with computer interfaces—reading screens, moving cursors, and performing actions on digital platforms. This feature has already shown promise in initial tests, with companies like Replit using it to automate multi-step software tasks.
The new Claude models and computer use beta are available now via the Anthropic API, Amazon Bedrock, and Google Cloud’s Vertex AI, opening up unprecedented possibilities for developers and businesses.