Today’s Anthropic news highlights Anthropic has released Claude Opus 4, a new AI model designed for software developers and engineers. The model achieved a 72.5% score on the SWE-bench benchmark, surpassing OpenAI’s GPT-4.1, which scored 54.6%. Opus 4 excels in complex problem-solving and can operate independently for up to seven hours, demonstrating sustained performance during rigorous tasks. It also features enhanced memory capabilities, allowing for better task awareness and coherence. The model is available via Anthropic’s API, Amazon Bedrock, and Google Cloud’s Vertex AI. ([itpro.com](https://www.itpro.com/software/development/anthropic-claude-opus-4-software-development?utm_source=openai)).
Anthropic Unveils Claude Opus 4, Outperforming OpenAI’s GPT-4.1
