Google Project Astra Is The Most Advanced AI Ever!?
Did Google just launch Project Astra? No — and the reason it feels that way is that two different companies are now using the same space word. (But did OpenAI take the name from Google previous launch in 2024!?) Google’s Project Astra is a research prototype first shown in 2024 (the bike-shop demo is from May 2025). OpenAI’s GPT-6 Astra is a separate, official model released on September 3, 2026. In this video I sort out which Astra is which, what is actually landing in Gemini right now, and why I think the least flashy — Qwen3.8-Omni-Flash deciding what NOT to watch — is the one that matters. 🔔 SUBSCRIBE TO BYTEFORWARD AI news, coding agents, model comparisons, benchmarks, and the things people are actually building: https://www.youtube.com/channel/UCGGHPOAzUaKsSc2yjsOeAFQ?sub_confirmation=1 So Google Project Astra needs your camera. Grok Bot needs your logins (and apparently your coffee budget). Which one do you wanna hand over first? 😂 Tell me in the comments. CHAPTERS 0:00 Google launches Project Astra?! 0:51 Two Astras: Google’s and OpenAI’s 1:12 What is actually new in Gemini 1:36 Both Astras on one leaderboard 1:58 Grok Bot can talk now 2:57 What Grok 2 promised in 2024 3:18 Qwen3.8-Omni-Flash: listen first, then look 4:25 Qwen vs Gemini: the actual numbers 5:21 Qwen2.5 said the plan out loud 5:44 GR-1: when seeing has to move an arm 6:41 Where I land 7:10 Your call SOURCES, IN THE ORDER THEY APPEAR Google — Project Astra demo film “Exploring the Capabilities of a Universal AI Assistant” (uploaded May 20, 2025): https://www.youtube.com/watch?v=JcDBFAm9PPI OpenAI on X — “This is GPT-6 Astra.” (Sep 3, 2026): https://x.com/OpenAI/status/2095595741528125780 OpenAI — GPT-6 Astra announcement: https://openai.com/index/gpt-6-astra/ Google DeepMind — Project Astra: https://deepmind.google/models/project-astra/ Google — Introducing Gemini 3.8 Live and 3.8 Live Extended Thinking (Sep 15, 2026): https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-8-live-gemini-3-8-live-extended-thinking/ Artificial Analysis on X — τ-Voice agentic benchmark (Sep 15, 2026): https://x.com/ArtificialAnlys/status/2099977684579434930 Grok Bot on X — “Grok Bot can talk now.” (Sep 17, 2026): https://x.com/bot/status/2100659463569170779 TestingCatalog on X — Grok Bot voice rollout (Sep 17, 2026): https://x.com/testingcatalog/status/2100677004286722385 SpaceXAI / xAI — Introducing Grok Bot (Aug 11, 2026): https://x.ai/news/introducing-grok-bot xAI — Grok-2 Beta Release (Aug 13, 2024): https://x.ai/news/grok-2 TestingCatalog on X — Qwen3.8-Omni-Flash (Sep 18, 2026): https://x.com/testingcatalog/status/2100852477725696424 Qwen on X — the demo video in that post is Qwen’s own, “Cases and more details” (Sep 18, 2026): https://x.com/Alibaba_Qwen/status/2100785971239534679 Qwen on X — “Performance and price” benchmark chart: https://x.com/Alibaba_Qwen/status/2100785967309496512 Qwen on X — Qwen3.8-Omni-Flash launch post: https://x.com/Alibaba_Qwen/status/2100785962414702599 Qwen — Qwen3.8-Omni-Flash blog: https://qwen.ai/blog?id=qwen3.8-omni-flash Alibaba Cloud Community — Qwen2.5: A Party of Foundation Models! (Nov 22, 2024): https://www.alibabacloud.com/blog/qwen2-5-a-party-of-foundation-models_601782 ByteDance Research — GR-1: Unleashing Large-Scale Video Generative Pre-training for Visual Robot Manipulation (2023): https://gr1-manipulation.github.io/ GR-1 paper: https://arxiv.org/abs/2312.13139 Footage and posts belong to their owners and are shown for commentary with on-screen credit. Narration, edit and captions are original; the ambient bed was made locally for ByteForward. #AI #ProjectAstra #Gemini #GPT6 #Qwen #Grok #Robotics
