Quantized Opus 5.5 says 'open' instead of 'run' — quantization quirks again
ryunuck · x · 2026-11-17
A user observed that after Anthropic quantized Opus 5.5 again, the model said 'open' for a command run instead of 'run' the command — 'open the command,' as if opening a ticket or filling out paperwork.
The post teases the subtle stylistic drift quantization can introduce into a model's language habits, noting this is a repeat occurrence after another quantization pass.
Related event: Quantized Opus 5.5 says 'open' instead of 'run' — quantization quirks again(2 posts)→
More from Models
- OpenAI urged to release rumored model allegedly cracking Navier–Stokes millennium problem — haider1 · 2026-11-17
- Verus Proof Index: new formal-verification benchmark evaluates 25 models' agentic proofs — fhuszar · 2026-11-17
- Asking ChatGPT and Claude to play NeurIPS reviewers — they reject the papers — burny_tech · 2026-11-17
- Benchmark cheating debate: scaling a 3B open model to ~60B explains the scores — RexDouglass · 2026-11-17
- Physician slams OpenEvidence's "lowering all-cause mortality" claim; Eric Topol amplifies — EricTopol · 2026-11-17
- wojak-bench: Community Arena Ranks AI Models by Whose 3D-Printed Figure Matches the Drawing Best — osanseviero · 2026-11-17