Agent Platform Scorecard

Who is building the agent operating system?

Twelve dimensions across seven providers. Re-graded as announcements land, with the market's reaction shown alongside.

xAI · Conversational UX

Natural interaction quality · updated 2026-06-01

Grok's voice mode and distinct personality make for engaging, fast interaction.

Field average: 7.1Dimension leader: OpenAI (9)

What this measures

How natural, fast, and trustworthy is the back-and-forth? Voice quality, latency, interruption handling, multimodality, and the sense that the assistant understood you the first time.

Across the web

Analysis
Grok 4 Launches With Benchmark Records and Idiosyncratic Behavior
The Batch (DeepLearning.AI) · 2025-07-16
Record reasoning/coding scores undercut by erratic, X-influenced behavior and protocol-skipping model changes.

Conversational UX across the field

OpenAI9
Google8
Microsoft7
Meta7
xAI7
Anthropic6
Apple6