New top AI video models, Google dominates, top 3D model generator, Grok 4.1, GPT-Codex-Max: AI NEWS

The video serves as a comprehensive news digest for AI enthusiasts, presenting a wide array of recent releases with visual demonstrations and benchmark comparisons. The host’s presentation is clear and well-structured, making complex topics accessible. The information is largely accurate and up-to-date, with references to official sources and GitHub repositories. However, the video lacks critical analysis; it primarily reports on the models’ capabilities without delving into potential limitations, biases, or ethical implications. The benchmarks cited are often from the model developers themselves, which may introduce bias. The host does not critically evaluate the claims or compare them with independent studies. Additionally, the video’s focus on speed and novelty may overshadow the need for rigorous scientific scrutiny. The inclusion of a sponsor segment is clearly marked and does not detract from the content. Overall, the video is a valuable resource for staying informed, but viewers should seek additional sources for a more balanced perspective.

The title accurately reflects the content, highlighting major AI releases including video models, Google's dominance, and notable model updates.

8/10

The video provides a comprehensive overview of recent AI releases, citing official sources and GitHub repositories. The information is presented with technical details and benchmarks, but lacks in-depth critical analysis and independent verification.

• 00:00AI news intro
• 01:13Depth Anything 3
• 03:46SAM3
• 05:48SAM3D
• 08:40HunyuanVideo 1.5
• 12:23Kandinsky 5.0
• 14:39Genspark
• 16:39Grok 4.1
• 17:21Gemini 3 and Nano Banana Pro
• 20:28GPT-5.1-Codex-MAX
• 22:10Proactive hearing assistant
• 24:18WeatherNext 2
• 25:52Google AntiGravity
• 27:43PhysX-Anything
• 29:39Dr Tulu
• 32:00Part X MLLM
• 33:42Uni-MoE 2 Omni

• Depth Anything 3 — Official project page for Depth Anything 3, a model for 3D mapping from images/videos.
• Segment Anything Model 3 (SAM 3) — Meta's blog post announcing SAM 3, an AI for object segmentation and tracking.
• SAM 3D — Meta's blog post about SAM 3D, a 3D model generator from images.
• HunyuanVideo 1.5 — GitHub repository for HunyuanVideo 1.5, an open-source video generation model.
• Kandinsky 5 — Official page for Kandinsky 5, an open-source video and image generation model.
• Grok 4.1 — xAI's announcement of Grok 4.1, a large language model.
• GPT-5.1-Codex-Max — OpenAI's page for GPT-5.1-Codex-Max, a coding model.
• Proactive Hearing Assistant — University of Washington project on a proactive hearing assistant.
• WeatherNext 2 — Google DeepMind's blog post about WeatherNext 2, a weather forecasting model.
• Google AntiGravity — Google's AntiGravity project page.
• PhysX-Anything — Project page for PhysX-Anything, a physics simulation model.
• Dr Tulu — GitHub repository for Dr Tulu, a deep research agent.
• Part X MLLM — Project page for Part X MLLM, a multimodal model.
• Uni-MoE 2 Omni — Project page for Uni-MoE 2 Omni, an omnimodal model.
• Gemini 3 review — Review video of Gemini 3 by AI Search.
• Nano Banana Pro review — Review video of Nano Banana Pro by AI Search.

• Depth Anything 3 — Official project page confirms the capabilities and benchmarks presented in the video.
• Segment Anything Model 3 (SAM 3) — Meta's blog post aligns with the video's description of SAM 3's features and performance.
• HunyuanVideo 1.5 — GitHub repository provides technical details and confirms the model's open-source availability.

• No discordant sources found — The video's claims are consistent with the cited official sources and repositories.

• aisearch.substack.com
• nvda.ws