Opus 5.2 spotted in Claude Code routing; leaked Model 2 reportedly writes most of Anthropic's code
新智元 · wechat · 2026-09-15
- Developers found Claude Code's "Opus5" silently routes to Opus 5.2 — apparently skipping 5.1. Reports: faster responses, cleaner output, and self-driven "gauntlet loops" that iterate code without laziness.
- A community "Tibo probe" (asking about an obscure figure with search off) distinguishes routed 5.2 sessions from old weights.
- A leaked August internal risk report claims an unreleased "Model 2" scored 62.8% on CoBench v2 (12.5 points above Mythos5) and already writes most of Anthropic's company code via agents; an RSI model reportedly scores 22 points higher still, said to replace 85% of research-team work.
Related event: Anthropic appears to be quietly testing Opus 5.2 via Claude Code(6 posts)→
More from Models
- Dev Lesson: You Can't Drop a Mid-Tier Model Into a Simple Task and Expect It to Figure It Out — rickasaurus · 2026-11-17
- Why AI should be avoided for sentence-level creative writing: everything becomes an explained symbol — zetalyrae · 2026-11-17
- Users Complain New Voice Mode Sounds Like an "Evil Millennial Tech Bro" — kaiden_yurk · 2026-11-17
- Same Prompt, Same Art: ChatGPT and Claude Produce Near-Identical Mathematical Fractals — burny_tech · 2026-11-17
- Devs: Gemini 4 Carbon is frontier for ML engineering, not deliberately crippled — olcan · 2026-11-17
- How do chat models distill thousands of tweets into a coherent latent persona? — teortaxesTex · 2026-11-17