Story: state of ai/leading models
Context: Vendor claim without contemporaneous independent adjudication in this record. Retrospectively, Artificial Analysis's current index (v4.1) estimates Grok 4 at 33 — above Gemini 2.5 Pro (26) among scored contemporaries — but the July 2025 frontier is only partially scored on today's index, so the superlative cannot be settled either way.