

GPT-5 vs Claude vs Gemini: 7 Brutal Real-life Tests
23 snips Aug 9, 2025
Uncover the truth about the highly anticipated GPT-5 as it's put to the test in real-world business scenarios. Discover how it measures up against Gemini 2.5 Pro and Claude Opus, revealing its strengths and critical flaws. The discussion highlights GPT-5's hallucinatory tendencies and surprises with the potential of GPT-5 Mini as a game-changer for small businesses. Learn about the transformative impact of a new, affordable API for developers, and see how AI can revolutionize tasks like email writing and ad analysis.
AI Snips
Chapters
Transcript
Episode notes
GPT‑5 Is A Major Upgrade
- GPT-5 is a major upgrade to ChatGPT that improves everyday usability for most users.
- Gael Breton ran 12 hours of hands-on tests and seven real-world scenarios to evaluate it.
Use Default GPT‑5; Reserve 'Thinking' For Complex Tasks
- Use the default GPT-5 for most queries and only switch to the 'thinking' mode for complex reasoning tasks.
- Monitor limits because thinking mode has stricter message quotas (200 messages per week).
GPT‑5 Hallucinated And Invented Features
- GPT-5 produced a detailed MCP integration guide that looked real but referenced features that don't exist.
- Gael Breton verified via web search and found the instructions were fabricated.