Benchmark ranks Grok, Muse, Dot, and Claude personal agents

A comparative benchmark evaluated four personal AI agents—Grok, Muse, Dot, and Claude—across nine tasks including Slack drafting, video editing, bug fixing, and inbox triage. Grok achieved a 93% completion rate and won three tasks: thumbnail creation, inbox triage, and shoe shopping. Muse was the fastest agent, completing four tasks quickest, and excelled in creative personal tasks like family newspapers and short-form video, though it failed tasks requiring business tool access like GitHub or Slack. Dot and Grok completed all nine tasks, while Muse and Claude completed only seven.

In coding, Dot and Claude produced the most accurate fixes for a Sentry bug, with Claude identifying an additional bug; Grok performed poorly by loosening validation. For Figma-based thumbnail design, Grok was rated the best. Claude produced the highest-quality launch video for a 'Product Intelligence' feature using 11 Labs MCP for voiceover, while Dot produced a low-quality PowerPoint-style output. The benchmark used operator agents to drive the other agents via computer use, with prompts kept to single-line, human-voice instructions. The evaluator argues that Grok is the best all-around personal agent for work tasks, Muse is superior for personal and creative life tasks, and Dot and Claude remain the top choices for coding workflows.