New DeepSeek & Gemini, Wan 2.5, Alibaba dominates AI, 3D scenes, new deepfake lipsync: AI NEWS

New DeepSeek & Gemini, Wan 2.5, Alibaba dominates AI, 3D scenes, new deepfake lipsync: AI NEWS

by AI Search

Applied Sciences & Engineering AI — News & Applications UYComputer scienceUYQArtificial intelligence (AI)
🎙 AI Search 👥 715K 📅 September 28, 2025 ⏱ 50 min 👁 103K 📄 news review 🧭 2026-08-03
Available in: English (current) Français

Keywords

AI newsDeepSeekGeminiWan 2.5Alibaba

Summary

This video is a weekly AI news roundup covering a wide range of recent developments. It starts with a humorous intro about the overwhelming pace of AI progress. The host then discusses several new AI tools and models: Video from 3D, which generates videos from a reference image and a 3D model; Lynx, an open-source reference-to-video generator from ByteDance; Hunyuan3D Part, a 3D model segmentation tool from Tencent; Lyra, an NVIDIA tool for generating 3D scenes from images or videos; DeepSeek 3.1 Terminus, an upgraded open-source model; OmniInsert, a tool for inserting characters or objects into videos; new Gemini 2.5 Flash and Flash-Lite models; a Unitree humanoid robot demo; a sponsored segment for Hailuo AI Agent; Skild’s adaptive brain for robots; Wan 2.5, a video generation model; and a series of Alibaba releases including Qwen 3 Max, Qwen 3 Omni, Qwen 3 VL, Qwen Image Edit 2509, and Qwen 3 Coder. The video also covers Kimi OK Computer, ChatGPT Pulse, Suno v5, and OmniHuman 1.5. The host provides demonstrations and comparisons, often highlighting improvements over previous models. The video concludes with a reminder of the rapid pace of AI development.

190 words

Critical Evaluation

The video serves as a comprehensive digest of recent AI developments, making it valuable for staying informed. The host demonstrates a good grasp of the technical aspects, explaining each tool’s functionality and potential applications. However, the analysis is largely descriptive rather than critical, often relying on vendor-provided benchmarks and demos without independent verification. The comparisons with previous models are useful but sometimes lack depth, as the host tends to accept improvements at face value. The sources cited are primarily official project pages and blog posts, which are credible but not always peer-reviewed. The video’s strength lies in its breadth and accessibility, but it could benefit from more critical scrutiny of the claims made. The adéquation titre/contenu is excellent, as the title accurately reflects the content. The presence of a sponsored segment is clearly disclosed and does not detract from the overall quality. The video’s tone is enthusiastic, which may appeal to a general audience but could be seen as overly promotional. Overall, the video provides a solid overview of the week’s AI news, but viewers should seek additional sources for deeper analysis.

182 words

Title / Content Match

The title accurately reflects the content, which covers multiple AI news items including DeepSeek, Gemini, Wan 2.5, and Alibaba's releases.

Quality & Reliability

7/10

The video provides a broad overview of recent AI releases and research, with links to primary sources. However, it lacks in-depth critical analysis and relies on vendor claims and benchmarks without independent verification. The presentation is engaging but sometimes hyperbolic.

Chapters

Cited Sources

  • Video from 3D — Project page for the Video from 3D tool.
  • Lynx — Project page for ByteDance's Lynx video generator.
  • Hunyuan3D Part — GitHub repository for Tencent's 3D model segmentation tool.
  • Lyra — NVIDIA research page for Lyra, a 3D scene generation tool.
  • DeepSeek 3.1 Terminus — Release notes for DeepSeek 3.1 Terminus.
  • OmniInsert — Project page for ByteDance's OmniInsert tool.
  • Gemini 2.5 Flash and Flash-Lite — Google blog post about the updated Gemini models.
  • Qwen 3 Max — Qwen blog post about Qwen 3 Max.
  • Qwen 3 Omni — Qwen blog post about Qwen 3 Omni.
  • Qwen 3 VL — Qwen blog post about Qwen 3 VL.
  • Qwen Image Edit 2509 — Qwen blog post about the image editing model.
  • Qwen 3 Coder — GitHub repository for Qwen 3 Coder.
  • Kimi OK Computer — Kimi's official website.
  • ChatGPT Pulse — OpenAI blog post about ChatGPT Pulse.
  • Suno v5 — Suno's official website.
  • OmniHuman 1.5 — Fal.ai model page for OmniHuman 1.5.

Concurring Sources

  • Artificial Analysis — Independent leaderboard ranking AI models, used to compare DeepSeek 3.1 Terminus.

External References

Contribution & Novelties

The video provides a timely and comprehensive roundup of recent AI developments, highlighting several open-source releases and research breakthroughs. It offers practical demonstrations and comparisons, making it accessible to a broad audience. The host’s enthusiasm and clear explanations help viewers grasp the significance of each tool.

Pour aller plus loin :

166 words

Radar Profile

The radar profile shows high scores in quantity of information and fiabilité globale, indicating a comprehensive and reliable news roundup. The quality of information and technical level are moderate, reflecting a balance between depth and accessibility. The overall profile suggests a well-rounded video suitable for staying updated on AI developments.

Reliability 7/10

💬 Très positif. Sur les 30 commentaires analysés, la majorité exprime une admiration pour la fréquence et la qualité des mises à jour, ainsi qu'un sentiment d'émerveillement face à la rapidité des progrès en IA. Certains commentaires mentionnent des références à des livres ou à des inquiétudes sur le rythme, mais le ton général est enthousiaste et reconnaissant.