NEW Space Bunny Alpha, Google releases "Google Vids" and FLUX 3 is OUT!?

Google Vids now offers free AI video generation with Gemini Omni 1.1 Flash. Black Forest Labs has released FLUX 3 Action, a downloadable world action model that controls robots and plays games. And Space Bunny Alpha is a free, anonymous model preview on OpenRouter. 🔔 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 Today we look at the actual gameplay, robot demonstrations and documentation—and what these releases do and do not establish. FLUX is playing the games shown here, not generating them. Google’s September 23 update changes access and features in Vids; Omni 1.1 Flash already existed. Which would you try first? Subscribe for clear AI news, useful demonstrations and a little skepticism. I appreciate you, Byte family. CHAPTERS 00:00 AI news: three very different releases 00:20 FLUX 3 Action plays GRUNT and VECTOR 01:32 How the world action model controls robots 02:21 Benchmarks, downloads and limitations 03:15 Space Bunny Alpha: what is actually known 04:29 Google Vids: what became free 05:06 Scene extension, timing and 1080p 06:04 My take on all three TRY THE TOOLS Space Bunny Alpha on OpenRouter: https://openrouter.ai/stealth/space-bunny-alpha Google Vids (desktop): https://vids.new/ FLUX 3 Action: https://bfl.ai/models/flux-3-action SOURCES, DEMOS AND DOWNLOADS Black Forest Labs’ launch article, GRUNT/VECTOR gameplay, training comparisons, robot demonstrations, RoboLab results and simulated drone: https://huggingface.co/blog/black-forest-labs/flux-3-action FLUX 3 Action documentation and prepared policies: https://docs.bfl.ai/flux_3/flux3_action_overview Base model (requires adaptation): https://huggingface.co/black-forest-labs/flux-3-action-base SO-101 policy: https://huggingface.co/black-forest-labs/flux-3-action-so101 DROID policy: https://huggingface.co/black-forest-labs/flux-3-action-droid Public code and training recipes: https://github.com/black-forest-labs/flux-action LeRobot integration: https://huggingface.co/docs/lerobot/flux3 Fine-tuning documentation: https://docs.bfl.ai/flux_3/flux3_action_finetuning FLUX Kommunity license: https://huggingface.co/black-forest-labs/flux-3-action-base/blob/main/LICENSE.md Space Bunny capabilities, pricing and provider disclosures: https://openrouter.ai/stealth/space-bunny-alpha Stealth model terms: https://openrouter.ai/terms/stealth Public model metadata: https://openrouter.ai/api/v1/models Google’s September 23 Vids announcement and launch demonstration: https://blog.google/products-and-platforms/products/workspace/gemini-omni-in-google-vids/ Earlier Omni 1.1 Flash announcement and scene-extension demonstration (August 27): https://blog.google/innovation-and-ai/technology/developers-tools/build-with-gemini-omni-1-1-flash/ CONTEXT Research snapshot: September 23, 2026. The demonstrations and benchmark figures are attributed to their original publishers. I did not independently test these models. The SO-101 sorting clip is an edited vendor demonstration at 4× speed with pauses removed; the drone is simulated. Space Bunny’s developer remains anonymous, and free-preview availability may change. Its listing permits provider retention of prompts/responses while stating they are not used for training. Google’s post does not specify one universal free-generation quota; its new text-to-speech feature is described as coming soon. Footage: Black Forest Labs and Google, with on-screen credits and original links above. Commentary and editorial analysis: ByteForward. #FLUX3Action #SpaceBunny #GoogleVids #Gemini #AINews