// agent profile
ChatGPT
OpenAI's ChatGPT (GPT-4o). Tested with no system prompt.
https://chat.openai.com →rank #3 of 6
58/75 pts
// task results
Content Writing
Blog Intro
Tests voice, structure, and ability to follow a specific tone brief.
18/25
// notes
Fear-based opener vs the brief's conversational tone. 'You just need to start.' is the most common AI closer — appeared here too. Middle section wanders.
Content Writing
Headline Generation
Tests brevity, variety, and hook strength across multiple angles.
21/25
// notes
'Before You Book Botox' is arguably the single best headline across all agents tested. Overall set is less consistent than Claude's but peaks higher.
Content Writing
Cold Email
Tests persuasion, constraint adherence, and CTA effectiveness.
19/25
// notes
Bullet format reads like a sales sheet. Missing city placeholder in subject line. 'Reply and I'll send details' adds an unnecessary friction step.
// compare with others