Anthropic's newest model, Claude Opus 5, is now available to the public, coming close to or exceeding the performance of its predecessor, Fable 5, in most areas for roughly half as much per token. Opus 5 outperforms Fable 5 on tasks like knowledge work, novel problem solving, and agentic search, but falls somewhat short in answering legal questions and performing multidisciplinary reasoning without additional tools, and is "substantially behind" Mythos 5 at exploiting cybersecurity vulnerabilities. During pre-deployment testing, Anthropic concluded Opus 5 showed lower rates of "misaligned behavior," lower rates of deceptive behavior, and proved safer in avoiding "reckless actions." This launch follows a period where Anthropic blocked access to Fable 5 for users worldwide after US government concerns about its offensive cybersecurity capabilities, with access restored several weeks later with restrictions. Anthropic is beta-testing automatic fallbacks, allowing users to reroute requests flagged by safety classifiers on Opus 5 or Fable 5 to another model. According to Sensor Tower's State of AI 2026 report, Claude's global market share was at 10.3% in May 2026, while ChatGPT's global share fell to 46.4% and Gemini hit 27.7%.