On Monday, Anthropic unveiled Claude Sonnet 4.5, a cutting-edge AI model excelling in coding benchmarks and designed for building “production-ready” applications. Available through the Claude API and chatbot, it retains the pricing of its predecessor at $3 per million input tokens and $15 per million output tokens. Developers and enterprises increasingly favor Anthropic’s models, with well-known companies like Apple and Meta utilizing them for software engineering tasks. Despite OpenAI’s GPT-5 outperforming Claude in some areas, Claude Sonnet 4.5 is said to offer superior coding performance, particularly for long-term projects, while also reducing sycophancy and vulnerability to prompt injection attacks. Additionally, Anthropic introduced the Claude Agent SDK to assist developers in creating custom agents and a temporary research preview, “Imagine with Claude,” for Max subscribers, showcasing real-time software generation. In an intensely competitive landscape, the rapid release pace emphasizes the need for continuous innovation in AI technology.
Source link
Share
Read more