TREE NEWS update: Anthropic’s Claude Opus 5.2 has quietly entered gray-scale testing inside Claude Code, developers found, with request status checks showing the backend model slug now points to Opus 5.2 even though the front-end name is unchanged. Users report that calling Opus 5 in Claude Code performs markedly differently from the web version. A prior internal risk report said Anthropic is deploying an unreleased ‘Model 2’ that writes most of the company’s code and would replace 85% of its research team.
Anthropic’s Claude Opus 5.2 Spotted in Gray-Scale Testing on Claude Code
The interesting signal here is not the version number but the delivery mechanism: a backend slug swap that leaves the user-facing name untouched means capability changes can reach production without any visible release event. That affects developers who treat the named model as a stable contract, and it complicates any external evaluation, since behavior differences between surfaces are now being noticed before any announcement. Whether gray-scale routing of this kind becomes standard across coding tools, and whether users can detect or pin a specific backend, is the open question worth watching.
Generated by AI for reference only.
Share on WeChat
Open WeChat → Scan → then tap "…" to send to a chat or Moments.
Tap "…" in the top-right corner to send to a chat or share to Moments.