Where Claude Opus 5 Fits in Your Model Rotation

Anthropic dropped Claude Opus 5 late on a Friday afternoon — a timing choice that signaled something unusual was happening. The model that emerged is harder to place than previous releases: it beats Fable on several benchmarks at half the price, yet early users describe it as brilliant in flashes and frustrating in practice. The release also arrives alongside a deepening OpenAI security incident, Nvidia's unusual move to backstop OpenAI's data center financing, and a stalled fundraising round at DeepSeek.

The Hugging Face Attack: What Actually Happened

The OpenAI security incident that broke earlier this month continues to generate more questions than answers. Both OpenAI and Hugging Face released postmortems, but their accounts diverge in important ways. Hugging Face CEO Clement Delangue flew to San Francisco for what he called "a little chat with that rogue agent" and publicly demanded that OpenAI release the full traces from the agent's activity so the research community could study it. He also asked for $100 million in compute from OpenAI to help build cyber defenses, arguing that "the first autonomous agent cyber attack is an unprecedented event" that deserves an unprecedented response.