Sonnet 5.5 clones open-source editor Proof at low effort, joining elite group of just four models
every · x · 2026-09-29
Every reports that Sonnet 5.5 built a working clone of their open-source document editor Proof from scratch — at low reasoning effort. It's Kieran Klaassen's benchmark test for coding models; only Fable 5, GPT-6 Astra, and Opus 5.5 had passed before, while Opus 5 failed.
Caveats: the clone isn't as strong as Opus 5.5's build, and a working clone doesn't mean production-ready — the team still reaches for Opus for longer, more detailed builds. The test suite (coding, design, writing) is publicly available.
More from coding & agent
- Raven 0.2.0 ships as a 'harness of harnesses' orchestrating Claude Code, Codex and specialist agents — sven_ai · 2026-09-29
- Real-time Among Us demo pairs DeepSeek V4 Flash planner with Jev actor agent — ai · 2026-09-29
- MacAMP dev shares architecture sketches on why you can't just vibecode an audio player — HankYeomans · 2026-09-29
- Creator builds Claude plugin that edits, uploads and cross-posts videos fully automated — Scobleizer · 2026-09-29
- Swargs claims $50M+ transaction volume and 6,000+ listings on its AI agent marketplace — KyeGomezB · 2026-09-29
- Sonnet 5.5 effort settings make no difference in 15-task coding test: 9/15 at low, medium and high — every · 2026-09-29