Artificial Analysis publishes AA-PersonalityBench: GPT-6 and Grok models are the most direct

Artificial Analysis published AA-PersonalityBench, a benchmark of model personality traits, with a results page, a methodology page and a tool for building your own personality ranking. The scores are the site's own.

Two of the traits it highlights:

• Direct (whether a model gets to the point without filler): GPT-6 Sol scores 100, Grok 4.6 98 and Grok 4.7 95.
• Answer First (whether a model leads with the answer before context, reasoning or caveats): Grok 4.5 scores 100, GPT-6 Sol 98 and GPT-6 Luna 95.

Artificial Analysis says GPT-6 and Grok models are the most Direct and most likely to use an Answer First format of all the models it measured.