Opus 5.5 shows only Opus-level chain-of-thought control, hinting size drives it
scaling01 · x · 2026-09-23
TheZvi observes that Opus 5.5 only has Opus-level ability to control its chain of thought, not Mythos-level, suggesting size and architecture drive this trait rather than overall capability. scaling01 counters that progress may have been made since Mythos 5.1. The exchange probes what actually determines chain-of-thought controllability, with no settled answer.
Related event: Opus 5.5's Chain-of-Thought Control Appears Scale-Dependent, Says TheZvi(2 posts)→
More from Models
- Stop benchmarking LLMs with 3D games, says Abacus.AI CEO — labs fine-tune for it — bindureddy · 2026-09-23
- Opus 5.5 impresses with a hand-crafted Mona Lisa SVG — djcows · 2026-09-23
- MazeBench 3D Spatial Benchmark: Opus 5.5 Hits 6% While GPT-6 Sol and Grok 4.7 Score Just 1% — daniel_mac8 · 2026-09-23
- MazeBench 3D Spatial Reasoning: Opus 5.5 Hits 6% While GPT-6 Sol and Grok 4.7 Manage Just 1% — patience_cave · 2026-09-23
- Opus 5.5 System Prompt Grows to 4,108 Words, Up From Opus 5's 3,225 — rajistics · 2026-09-23
- Hand-edited poem still flagged 100% AI by Pangram — detector false positive debate — JFPuget · 2026-09-23