How to Prompt Each AI Model
ChatGPT, Claude, and Gemini are built differently โ and respond better to different prompting approaches. These guides cover exactly what changes between models and why.
ChatGPT (GPT-4o)
ChatGPT responds well to direct instructions, tolerates ambiguity better than most models, and performs particularly well on structured tasks. But it has specific failure modes that different prompting techniques can prevent.
Best for
- โInstruction-following on complex, multi-step structured tasks
- โCode generation and debugging across most languages
- โDocument summarisation and data extraction
Claude (Anthropic)
Claude is designed to be honest, to push back on flawed premises, and to reason through problems rather than pattern-match to an expected answer. These traits make it different to work with โ and often produce more reliable outputs when prompted correctly.
Best for
- โLong-document analysis (200k+ token context window)
- โNuanced writing tasks requiring judgment โ tone, subtext, sensitivity
- โStructured reasoning and analysis where accuracy matters
Gemini (Google)
Gemini is Google's flagship AI model, tightly integrated with Google's ecosystem. It has specific strengths around real-time information access, multimodal tasks, and Google Workspace integration โ and specific prompting patterns that unlock those strengths.
Best for
- โReal-time information via Google Search grounding
- โGoogle Workspace integration (Docs, Gmail, Sheets, Slides)
- โMultimodal tasks โ images, PDFs, and text in the same prompt
Quick Comparison
| Use Case | ChatGPT | Claude | Gemini |
|---|---|---|---|
| Complex reasoning | โ โ โ โ | โ โ โ โ โ | โ โ โ โ |
| Creative writing | โ โ โ โ โ | โ โ โ โ | โ โ โ |
| Code generation | โ โ โ โ โ | โ โ โ โ | โ โ โ โ |
| Long document analysis | โ โ โ | โ โ โ โ โ | โ โ โ โ โ |
| Real-time information | โ โ โ | โ โ | โ โ โ โ โ |
| Instruction following | โ โ โ โ โ | โ โ โ โ โ | โ โ โ โ |
| Google Workspace integration | โ | โ | โ โ โ โ โ |
Ratings are relative, not absolute. All three models are capable across most tasks.
Working with Claude?
If Claude is your main model, start with the Claude Code guide โ the most practical skill upgrade for writers, marketers, and operators in 2026.
Read the Claude Code guide