Gemini Canvas vs GPT-5: Which Creates the Best Presentation?

Introduction

Preparing a clear and aesthetic presentation can quickly become time-consuming. With Canvas, its new feature integrated into Gemini, Google promises to transform simple text or a URL into a complete and visual presentation.

I wanted to test this promise against GPT-5, on a concrete case:

Create a Google Slides presentation from my article: 👉 Developer Skills Evolution in 2025

The exact prompt I used in both AIs:

🧠 Prompt: “I want to create a Google Slides presentation on https://nicolas-dabene.fr/articles/2025/11/10/evolution-competences-developpeurs/”

The results are very different… and revealing.

🎨 Google Gemini’s Canvas: Fluid and Automated Creation

Gemini introduces a true integrated visual creation studio. Thanks to Canvas, any user can generate a slide deck from text, a link, or even uploaded images.

🔧 How It Works

  1. Open Gemini → Tools → Canvas section
  2. Paste your topic, text, or URL
  3. Click on Create presentation
  4. Gemini automatically generates:
    • A structured plan (introduction, content, conclusion)
    • Coherent themes with colors and typography
    • AI-generated images
    • And even LaTeX if needed for equations or tables

💡 Canvas allows you to edit content directly, add slides, then export to Google Slides with one click.

🎯 What I Got With My Prompt

In a few seconds, Gemini generated:

  • Coherent and well-hierarchized slides
  • A clean visual identity, based on Google Slides tones
  • Relevant visuals evoking the future of development and AI
  • A logical plan:
    1. Introduction to context
    2. Skills mutation
    3. New market expectations
    4. AI’s role in transformation
    5. Conclusion and perspectives

The result is clean, readable and immediately usable.

🤖 GPT-5: A Disappointing and Unusable Result

For its part, GPT-5 produced a result significantly below Gemini Canvas. Without a very elaborate and structured prompt upfront, the generated presentation is not directly usable.

Main Weaknesses

  • Unusable raw output: requires major rework to be presentable
  • No native visual generation (requires a plugin for images)
  • Incoherent structure without a very precise and detailed prompt
  • Absent or haphazard layout depending on generations
  • Often requires multiple iterations to get something correct

What GPT-5 Can Do (With Much Effort)

  • Possible integration with PowerPoint or Google Slides via API (but complex setup)
  • Interesting narrative potential if and only if the prompt is very well crafted
  • Can generate dense content, but often too verbose for slides

⚠️ Verdict: For presentation generation, GPT-5 is not up to par with Gemini Canvas. The time saved by automation is lost in rework and adjustments.

⚔️ The Synthetic Comparison

Criterion Gemini Canvas GPT-5
Ease of use ⭐⭐⭐⭐⭐ ⭐
Visual style Clean, professional Non-existent without rework
Content richness Medium but usable Too dense, poorly structured
Plan creativity Basic but effective Incoherent without elaborate prompt
Slides integration Native (1 click) Via API (complex)
LaTeX / PDF support Yes Partial
Immediate result ✅ Directly usable ❌ Requires major rework
Target audience All users Experts willing to iterate extensively

📚 Some Inspiring Use Cases

Gemini Canvas is not limited to presentations:

  • 🎓 Students: transform revision sheets into interactive quizzes
  • 💼 Professionals: create action plans from meeting notes
  • 🧑‍💻 Developers: visualize code, logs or AI results

And thanks to integrated LaTeX rendering, scientists can now prepare complex equations without additional software.

📂 Test Results

I put the two presentations side by side — Gemini vs GPT-5 — so you can judge for yourself:

➡️ Download the presentations:

👉 Each presentation uses the same base, but tone, layout and content density differ radically.

Conclusion

Between Gemini Canvas and GPT-5, the verdict is clear: Gemini Canvas wins by a wide margin for presentation creation.

  • Gemini Canvas delivers an immediately usable result, with visual coherence, clear structure, and native integration with Google Slides. It’s the ideal tool for anyone looking to save time.

  • GPT-5, on the other hand, disappoints on this task: without a very elaborate prompt upfront, the result requires significant rework that cancels out the expected productivity gain. It is not suited for visual presentation generation.

My recommendation: for creating presentations quickly and effectively, Gemini Canvas is the obvious choice. GPT-5 might occasionally serve as a complement to refine textual content, but certainly not as a primary slide creation tool.

Article published November 14, 2025 by Nicolas Dabène – AI Expert & PrestaShop Developer