Je teste Gemini 1.5 Pro — CHOQUÉ ! (Google a tout cassé)

Je teste Gemini 1.5 Pro — CHOQUÉ ! (Google a tout cassé)

by Ludo Salenne

Applied Sciences & Engineering AI — Models & Research UYQArtificial intelligence (AI)UYQFGenerative artificial intelligence
🎙 Ludo Salenne 👥 267K 📅 March 27, 2024 ⏱ 27 min 👁 59K 📄 expert opinion 🧭 2026-08-21
Available in: English (current) Français

Keywords

Gemini 1.5 ProGoogle AImultimodal1 million tokensAI test

Summary

In this video, Ludo Salenne tests Google’s Gemini 1.5 Pro, an AI model with a 1 million token context window. He replicates several official Google demos to verify their authenticity. The tests include analyzing a 8-minute video summary of the 1998 World Cup final, where Gemini correctly identifies the first goal at 1:09 and the scorer (Zinedine Zidane). He then tests multimodal understanding by providing a drawing and asking where it appears in the video, which Gemini finds at 6:54. Next, he uploads a longer YouTube video and asks for optimized titles; Gemini accurately suggests titles related to Abercrombie & Fitch’s marketing decline, despite not having access to audio. He then tests PDF analysis with his 217-page book, successfully retrieving the table of contents, identifying the most repeated phrase, and describing a specific burnout event with precise details. Finally, he tests the folder upload feature with five white papers, asking questions across the documents. The creator is impressed by the model’s performance, especially its ability to understand context and retrieve specific information from large files. He concludes that Gemini 1.5 Pro is a significant advancement in AI, potentially surpassing competitors.

190 words

Critical Evaluation

Value of the Information & Strength of the Argument

The video provides valuable hands-on demonstrations of Gemini 1.5 Pro’s capabilities, particularly its large context window and multimodal understanding. The creator systematically replicates official demos, adding credibility to his assessment. He also tests edge cases, such as finding a deliberately inserted sentence in a PDF, which showcases the model’s deep comprehension. The argumentation is persuasive, but it is based on anecdotal evidence and personal enthusiasm rather than rigorous benchmarking.

Scientific Rigor, Source Quality, Title Accuracy

The creator references Google’s official demos and provides links to them in the description. He also links to the AI Studio platform for viewers to test the model themselves. However, he does not cite independent studies or comparative analyses. The title accurately reflects the content, and the video’s structure is clear with chapters. The creator’s methodology is transparent, but the lack of external validation limits the scientific rigor.

152 words

Title / Content Match

The title accurately reflects the content: the creator tests Gemini 1.5 Pro and expresses astonishment at its capabilities.

Quality & Reliability

6/10

The video is a hands-on test of Gemini 1.5 Pro, replicating Google's official demos. The creator provides transparent methodology and verifiable results, but the assessment is subjective and lacks independent verification or comparison with other models.

Chapters

Cited Sources

Concurring Sources

  • Google AI Studio — The platform used for testing, confirming the model's availability.

External References

Contribution & Novelties

The video offers a practical, user-level evaluation of Gemini 1.5 Pro, highlighting its real-world applications for video analysis, document processing, and multimodal understanding. It goes beyond marketing claims by replicating demos and testing edge cases, providing valuable insights for potential users.

Pour aller plus loin :

77 words

Radar Profile

The radar profile shows moderate scores across all dimensions, with a slight emphasis on information quantity and reliability. This reflects a video that is informative and generally credible but lacks deep technical analysis and external validation.

Reliability 6/10

💬 Très positif. Sur les 30 commentaires analysés, la grande majorité exprime un enthousiasme marqué pour les capacités de Gemini 1.5 Pro et remercie le créateur pour sa démonstration, certains mentionnant l'impatience de tester l'outil.