News Summary

Anthropic has launched its latest AI model, Claude Opus 4.5, which the company claims is its most capable system to date. Positioned as the “best model in the world” for coding, agents, and computer-based tasks, Opus 4.5 demonstrates significant advancements in software engineering, outperforming other models on benchmarks testing real-world coding problems. Available through Anthropic’s apps, API, and major cloud platforms, the model is priced at $5 US per million input tokens and $25 for output tokens. Beyond coding, Opus 4.5 excels in research tasks and document handling, with improved reasoning and agent capabilities. The model’s enhancements include better handling of ambiguous instructions and multi-system bugs, while also offering stronger safety measures against prompt-injection attacks. These developments suggest potential shifts in various professions, including engineering, due to AI’s evolving capabilities.