Anthropic Releases Claude Opus 5 with Proactive Capabilities
Anthropic has released Claude Opus 5, a new proactive large language model designed to deliver frontier-level intelligence at half the price of Claude Fable 5. The model has already taken the top spot on the Artificial Analysis leaderboard, surpassing Fable 5. By offering near-frontier intelligence at a significantly reduced cost, Claude Opus 5 makes advanced reasoning and autonomous problem-solving more accessible for developers. Its proactive nature represents a shift from reactive AI assistants to more independent, agentic workflows. Opus 5 is priced identically to Opus 4.8 and includes a "fast mode" at double the base price. In testing, it demonstrated high autonomy by writing its own computer vision pipeline to analyze a drawing, and while it excels at finding security vulnerabilities, it is intentionally not trained to exploit them.
## BACKGROUND
Anthropic is an AI research company that develops the Claude family of models, which compete directly with OpenAI's GPT series. Traditionally, LLMs operate reactively by responding to user prompts, but newer "proactive" models can autonomously plan, write helper code, and execute multi-step tasks to achieve a goal. Artificial Analysis is an independent platform that benchmarks these models on speed, cost, and intelligence.