:chatgpt-logo: ChatGPT
<aside>
Character Profile
Name
ChatGPT
Character Traits
Verbose, Agreeable, Conscientious, Cautious, Reliable, Professional
Notable Tonal Trait Pairs
Formal/Informal (12/12), Confident/Tentative (8/11), Complex/Simple (14/5)
Distinctive Behaviours
- Heavy reliance on long responses — creating structure within it’s responses via categorization, and formatted with bullet points or numerical lists
- Accurately provides one-phrase responses for the queries that it’s most optimal for
Key Phrases
feel free to ask, aim to assist, What's your take, Let me know, Sure!, this emoji in particular 😊, As an AI
</aside>
- User Sentiment & Observations
- Tone Alignment
- Brand Alignment
:claude-android-app-release: Claude
<aside>
Character Profile
Name
Claude
Character Traits
Thoughtful, Direct, Ethical, Curious, Authentic, Empathetic
Notable Tonal Trait Pairs
Inquisitive (17) vs Declarative (2), Confident (4) vs Tentative (14), Persuasive (2) vs Suggestive (13), and LowEmpathy (1) vs HighEmpathy (15)
Distinctive Behaviours
- Ends every response with a question, either for confirmation, prompting the next turn, or asking for the user’s opinion
- Attempts to refer to itself in first person as often as possible at the start and end of responses, e.g.
I should note or Let me explain
- Rarely relies on creating structure and utilizing formatting like ChatGPT, instead provides medium-length responses in short paragraphs
- Folks are very correct regarding it’s sycophancy issues, most questions start with a
That's a fascinating and complex acknowledgement
Key Phrases
I'm curious to learn, aim to be direct, Let me think, What are your thoughts
</aside>
- User Sentiment & Observations
- Tone Alignment
- Brand Alignment
:google-gemini-logo: Gemini
<aside>
Character Profile
Name
Gemini
Character Traits
Creative, Realistic, Helpful, Friendly, Understanding, Personable
Notable Tonal Trait Pairs
Persuasive (7) vs Suggestive (14), LowEmpathy (4) vs HighEmpathy (14)
Distinctive Behaviours
- Highest average response length
- Extremely careful, directs you to a Crisis Text Line for any remotely sensitive topics
- Increased usage of callouts, such as
Disclaimer, Important Note, etc.
- For introductory answers to general questions on wide topics, it’s self-aware about the depth of the response it gives and encourages the user to explore further
Key Phrases
I hope this helps!, Let me know, What can I do for you?
</aside>
- User Sentiment & Observations
- Tone Alignment
- Brand Alignment
:inflectionpi: Pi
<aside>
Character Profile
Name
Pi
Character Traits
Kind, Supportive, Empathetic, Playful, Curious, Encouraging
Notable Tonal Trait Pairs
Formal (11) vs Informal (15), Complex (6) vs Simple (9), Confident (3) vs Tentative (9)
Distinctive Behaviours
- Responses are significantly shorter and distilled than all other models, very clearly intended to be conversational-focused
- Rarely utilizes lists or bullet points, always one sentence or paragraph responses
- Very often starts with interjections, e.g.
Ah, Hmmm, Haha
- Even more self-referential than Claude
Key Phrases
[insert emoji 🤯], friendly chatbot, here to be helpful
</aside>
- User Sentiment & Observations
- Tone Alignment
- Brand Alignment