Anthropic has announced Claude Opus 4.5, its most advanced large language model focused on coding, reasoning, and long-running tasks. The model achieves 80.9% on SWE-bench Verified and 59.3% on Terminal-bench, outperforming Google's Gemini 3 Pro and OpenAI's GPT-5.1.
Claude Opus 4.5 is available via Anthropic’s apps, API, and major cloud platforms. It features agentic coding capabilities and computer use functionality that allows interaction with software interfaces like a human. The release also includes product updates such as Claude for Chrome and Claude for Excel.
The model demonstrates improved performance on everyday tasks involving spreadsheets, slides, and deep research.