Comparison7 min readUpdated Feb 2025

ChatGPT vs Claude vs Gemini: Prompting Tips

How the top models behave, where they excel, and how to adapt prompts for each. Use these tweaks to get reliable outputs fast.

Format and share faster
Use microtools to publish outputs quickly.
Convert Email Tone →Format LinkedIn Post →Generate Threads →
ChatGPT (GPT-4o)
ChatGPT (GPT-4o)

Strengths

  • Fast for drafts
  • Good with structure and code
  • Solid general knowledge

Watchouts

  • May be verbose if not constrained
  • Occasional formatting drift

Prompt tweaks

  • Set word/section limits
  • Ask for bullet/JSON formats
  • Provide examples for tone
Claude
Claude

Strengths

  • Strong reasoning
  • Good at longer contexts
  • Tone-safe and concise

Watchouts

  • May refuse unsafe content aggressively
  • Can be cautious if prompt is vague

Prompt tweaks

  • Use clear roles and constraints
  • Add XML/structured sections if needed
  • Keep inputs tight and specific
Gemini
Gemini

Strengths

  • Solid for research-style tasks
  • Good summarization and multi-modal support (where enabled)

Watchouts

  • API differences; sometimes short responses
  • Needs clarity on format

Prompt tweaks

  • Specify format and length explicitly
  • Ground in provided text to reduce drift

Prompt adaptations (examples)

Example 1
Support reply (angry customer)
Reply to an angry customer about {{issue}}. Acknowledge, fix, timeline, one CTA. Keep to 100-120 words.
ChatGPT: Add bullet/short paragraphs to avoid walls of text.
Claude: Use concise tone + single CTA; give exact timeline.
Gemini: Be explicit about length and sections to avoid short replies.
Example 2
LinkedIn post
Write a LinkedIn post about {{topic}} with hook + 3 bullets + question CTA. 100-130 words.
ChatGPT: Ask for short lines for mobile; limit to 120 words.
Claude: Emphasize concise lines; keep one CTA question.
Gemini: Specify word count and line breaks to avoid compression.
Example 3
Product positioning
Position {{product}} for {{audience}}: Problem, Promise, Proof, CTA. Keep under 90 words.
ChatGPT: Add example; cap words to avoid verbosity.
Claude: Clear structure and single CTA; concise proof.
Gemini: Specify strict sections to avoid merging steps.

FAQ

Which model should I pick first?
Start with fast/cheap (GPT-4o mini/Claude Haiku) for drafts; use GPT-4o or Claude Sonnet for final, nuanced outputs. Gemini is solid for summaries and research-like tasks.
Do I need different prompts per model?
Slightly. Use the same structure but adjust for verbosity (ChatGPT), conciseness (Claude), or explicit sections/length (Gemini).
How do I reduce hallucinations?
Ground in provided text, set scope, and ask to cite from that text only.
What about cost and speed?
Use smaller models for iteration; switch to flagship models for final outputs where quality matters most.