¡Volvió Deepseek v4! Al nivel de GPT y Claude y 10 veces más barato

¡Volvió Deepseek v4! Al nivel de GPT y Claude y 10 veces más barato

🎙 EDteam 👥 1.0M 📅 April 30, 2026 ⏱ 24 min 👁 157K 📄 expert opinion 🧭 2026-08-02
Available in: English (current) Français

Keywords

DeepSeek V4open sourceHuaweibenchmarkspricing

Summary

The video, presented by EDteam, discusses the release of DeepSeek V4, a new open-source AI model from China. It provides historical context, comparing it to previous DeepSeek models and the impact of DeepSeek R1 in January 2025. The presenter highlights that DeepSeek V4 was trained using Huawei chips, marking a significant step towards China’s independence from Nvidia hardware. The video compares DeepSeek V4 Pro and Flash versions, detailing their parameter counts and pricing, which is significantly lower than GPT-5.4 and Claude Opus 4.6. It includes benchmark results showing DeepSeek V4 performing competitively in various tests, though slightly behind the latest models. The presenter conducts two real-world tests: creating an interactive dashboard and a commercial presentation, comparing DeepSeek’s performance against ChatGPT and Claude. The conclusion emphasizes the importance of cost-effectiveness and encourages viewers to focus on creating rather than brand loyalty. The video also promotes EDteam’s courses and scholarships.

148 words

Critical Evaluation

The video provides a comprehensive and accessible overview of DeepSeek V4, situating it within the broader context of the AI race between the US and China. The presenter, Álvaro, demonstrates a strong understanding of the technical aspects, such as mixture of experts, parameter counts, and pricing structures. He effectively explains complex concepts like active parameters and context windows using analogies, making the content understandable for a general audience. The historical background, including the impact of DeepSeek R1 and the role of Huawei chips, adds valuable context. The comparison of benchmarks and pricing is well-structured and informative, highlighting DeepSeek’s cost advantage. However, the video has some limitations. The presenter’s analysis is partly subjective, particularly in the real-world tests, where he acknowledges potential biases. The claim that DeepSeek V4 is ‘10 times cheaper’ is based on API pricing, but the actual cost savings depend on usage patterns. Additionally, the video does not provide direct citations for the benchmark data or pricing, which reduces its scientific rigor. The presenter also notes that DeepSeek V4 is slightly behind the latest models from OpenAI and Anthropic, which tempers the enthusiasm. Overall, the video is a valuable resource for understanding DeepSeek V4’s position in the market, but viewers should seek additional sources for verification. The adéquation between title and content is good, as the video indeed discusses DeepSeek V4’s capabilities and pricing. The video’s strength lies in its clear explanations and practical demonstrations, making it a useful introduction to the topic.

245 words

Title / Content Match

The title accurately reflects the content: the video discusses DeepSeek V4's capabilities, pricing, and its competitive position against GPT and Claude.

Quality & Reliability

7/10

The video provides a balanced overview of DeepSeek V4, including benchmarks, pricing, and context on Chinese AI development. The creator demonstrates technical knowledge and includes real-world tests, but some claims lack direct citations and the analysis is partly subjective.

Chapters

Cited Sources

Concurring Sources

Dissenting Sources

  • Independent benchmark analysis — Some independent analyses may show different benchmark results or pricing comparisons, but no specific source is provided in the video.

External References

Contribution & Novelties

The video provides an up-to-date analysis of DeepSeek V4, highlighting its open-source nature, competitive pricing, and the significance of being trained on Huawei chips. It offers a practical comparison with leading models and contextualizes the development within the US-China tech rivalry.

Pour aller plus loin :

94 words

Radar Profile

The radar profile shows strong scores in quantity of information and technical level, indicating a content-rich video with good depth. Quality and reliability are slightly lower, reflecting the subjective nature of some claims and lack of direct citations.

Reliability 7/10

💬 Positif. Sur les 30 commentaires analysés, la majorité exprime une appréciation positive, soulignant l'utilité de la comparaison, la performance de DeepSeek et le message final motivant.