Anthropic has released Claude 4, a large language model that the company claims surpasses OpenAI's GPT-4 in reasoning and reliability. The model is the first AI system to achieve a perfect score on the bar exam and demonstrates unprecedented capabilities in complex problem-solving.

The release intensifies competition in the AI industry, where OpenAI has maintained a lead since ChatGPT's launch in 2022. Anthropic, founded by former OpenAI researchers, has positioned itself as focused on AI safety while still pushing the frontier of capabilities.

Enterprise customers including Amazon, Notion, and Quora have already integrated Claude 4 into their products.