NEW GPT-4.1 DESTROYS Gemini 2.5 & Claude? 🤯

Key Concepts:

• GPT-4.1: OpenAI's advanced language model.
• Claude: An AI assistant developed by Anthropic.
• Gemini 2.5 Pro: Google's AI model, part of the Gemini family.
• Open Router: A platform to access various AI models via API.
• AI Detectors: Tools to identify AI-generated content (e.g., ZeroGPT).
• P5.js: A JavaScript library for creative coding.
• LiveWeave: A real-time HTML, CSS, and JavaScript editor.
• AI Profit Boardroom: A community and resource hub for AI-related topics.

1. Content Creation Test

• Prompt: Write a humorous 1500-word blog post on "Google's next algorithm will ruin your life or not."
• GPT-4.1:
  • Fastest response time.
  • Content was highly humanized and natural.
  • AI detection score: 0% (undetectable by ZeroGPT).
• Claude:
  • Content was good but slightly over the top.
  • AI detection score: 10.96%.
• Gemini 2.5 Pro:
  • Content was less relevant and went off on tangents.
  • AI detection score: 1.89%.
• Result: GPT-4.1 won in terms of speed, content quality, and humanization.

2. Coding Task 1: Pixelated Dinosaur Game (P5.js)

• Prompt: Create a pixelated dinosaur game using P5.js.
• GPT-4.1:
  • Generated functional game code.
  • Minor UI issues (green on green).
  • Game worked perfectly.
• Gemini 2.5 Pro:
  • Buggy output; the game didn't work properly.
  • The presenter noted that earlier versions of Gemini 2.5 Pro performed better on this task.
• Claude:
  • Generated non-functional code (total trash).
• Result: GPT-4.1 won due to the functional game output. Gemini came in second because it did actually work unlike Claude's output.

3. Coding Task 2: Interactive Water Molecule Simulation (HTML, CSS, JavaScript)

• Prompt: Create an interactive water molecule simulation using HTML, CSS, and JavaScript.
• GPT-4.1:
  • Generated code that didn't work at all (terrible output).
• Gemini 2.5 Pro:
  • Generated a functional simulation with temperature controls.
  • Lacked a UI feature (simulation legend).
• Claude:
  • Generated a functional simulation with a simulation legend.
• Result: Claude won due to the functional simulation and UI legend, followed by Gemini 2.5 Pro. GPT-4.1 came in last.

4. Coding Task 3: SEO Calculator Landing Page (HTML)

• Prompt: Create a calculator landing page for the niche SEO calculators as a single HTML file.
• GPT-4.1:
  • Fastest response time.
  • The UI was not great and not responsive.
  • The pop-up couldn't be closed.
• Claude:
  • The UI was weird and the box didn't work.
• Gemini 2.5 Pro:
  • Generated a better landing page UI.
  • Failed to build out the SEO calculator tool.
• Result: All models performed poorly. Claude was ranked first, GPT-4.1 second, and Gemini 2.5 Pro last.

5. General Observations and Opinions

• GPT-4.1 is the fastest in generating responses.
• AI models tend to perform best on the first day of release and then get dialed down.
• GPT-4.1 is a paid model, while Claude and Gemini 2.5 Pro can be used for free.

6. Final Verdict

• Coding Tasks: Claude
• Writing Tasks: GPT-4.1
• Free Option: Gemini 2.5 Pro

7. Resources Mentioned

• AI Profit Boardroom (for prompts and AI learning resources)
• Windurf (for free 7-day access to GPT-4.1)
• astudio.google.com (for free access to Gemini 2.5 Pro)
• editor.p5js.org (P5.js editor)
• LiveWeave (HTML, CSS, JavaScript editor)