Arena's Alignment Index: OpenAI sweeps all seven top spots, Opus 5.5 falls to #17

BorisMPower · x · 2026-11-22

LMArena updated its Alignment Index, and OpenAI now holds all seven top spots. Fable 5.2 debuts at #8, while Anthropic's Opus 5.5 drops from #6 to #17 under the new methodology. The best open model is Kimi K3.1 Pro at #9.

The index measures how agents behave in real sessions: doing things you didn't approve, claiming work is done when it isn't, and crediting you with things you never said. The poster's takeaway: Claude may still write the best code, but on behaving, nobody is close to OpenAI.

Related event: LMArena Alignment Index update: OpenAI sweeps top seven, Kimi K3.1 Pro jumps to 9th(6 posts)→

Original post →

More from Models

Models channel →