=== AI__TSUBAKI in-window originals (credit table) === Fri Nov 13 04:45:23 +0000 2026 3L/230V media=['video'] This is native 4K. You can take a Draft straight to 4K, so the cost stays low. Seedance 2.5 credits for 10s on @dreamina_ai 480p: 220 720p: 480 1080p: 1170 4K: 2260 #DreaminaCPP https://t.co/xnm6slejB1 https://x.com/AI__TSUBAKI/status/2121102100142977057 Fri Nov 13 03:57:13 +0000 2026 0L/87V media=['photo'] InkoAI Agent, the new creative assistant from DomoAI, is officially live. From one sentence to a post-ready video, all in a single chat. #DomoAI #InkoAIAgent https://t.co/bUkX5pDo4m https://x.com/AI__TSUBAKI/status/2121089977505493423 Fri Nov 13 01:03:49 +0000 2026 18L/809V media=['video'] First try with Seedance 2.5 in 4K. It's clean, for sure. Honestly, though, the jump felt smaller than I expected. Still beautiful, no question. I'll keep testing, including a side-by-side with 1080p. Made with Seedance 2.5 on @openart_ai #OpenArt https://t.co/YAgSbdPMcX https://x.com/AI__TSUBAKI/status/2121046340184735941 Thu Nov 12 00:16:01 +0000 2026 134L/5615V media=['video'] AI video has no concept of LOG. The color comes out baked in, so there's only so far you can push it in the grade. To get closer to the look at generation time, I found a film still on FILMGRAB and used it as a color-only reference. Made with Seedance 2.5 on @dreamina_ai #DreaminaCPP https://x.com/AI__TSUBAKI/status/2120671924640727307 Wed Nov 11 21:27:31 +0000 2026 35L/2555V media=['video'] Days of care, then this Akari and her Tamagotchi Made with Seedance 2.5 on @openart_ai #OpenArt https://t.co/mv87zqL5bL https://x.com/AI__TSUBAKI/status/2120629519665316094 Wed Nov 11 20:29:41 +0000 2026 4L/342V media=['video'] Riding like this must feel amazing... Winter is almost here! Seedance 2.5 Draft → Smart HD Made with @nadou_pro https://t.co/vKzf6Uf5Z8 https://x.com/AI__TSUBAKI/status/2120614963421343985 Wed Nov 11 09:41:07 +0000 2026 11L/605V media=['video'] Just checking her hair That's it, that's the video Seedance 2.5 Draft → Final 1080p Made on @renoiseai @renoiseaijp https://t.co/af9kxl4AXE https://x.com/AI__TSUBAKI/status/2120451746959135142 Wed Nov 11 06:39:25 +0000 2026 6L/425V media=['video'] Freezing room + growling stomach = 3 minutes to fix both Made with Dreamina Marketing Studio on @dreamina_ai It brings professional-grade, high-aesthetic marketing visuals to life with impressive efficiency. #DreaminaMarketingStudio #DreaminaAI https://t.co/3r7iB6UayR https://x.com/AI__TSUBAKI/status/2120406021919408530 Tue Nov 10 07:56:09 +0000 2026 8L/612V media=['video'] 7-5-3: Shichi-Go-San Long ago in Japan, so many children died young that a child was said to belong to the gods until the age of 7. Odd numbers were considered lucky, so families gave thanks at 3, 5 and 7. Every November, kids still dress in kimono and visit a shrine. Made with PixVerse V6 Max on @wavespeed_ai https://x.com/AI__TSUBAKI/status/2120062943278874897 Tue Nov 10 00:28:23 +0000 2026 9L/1315V media=['video'] Ninjas Don't Need Stairs Made with Pollo Recast on @itsPolloAI @polloai_creator #PolloAI https://t.co/mKNEveGBpf https://x.com/AI__TSUBAKI/status/2119950261787717712 === STANDING OFFER STATUS GREP (in-window originals) === -- DaVinci: 0 -- Topaz: 1 @CharaspowerAI Wed Nov 11 12:00 19L/1792V | Being able to push AI-generated videos into 4K HDR with Topaz Astra clearly changes the game. 🔥 It’s not just about maki -- Topview: 0 -- Nebius: 0 -- Magnific: 0 -- Kling_4_unlimited: 0 -- Genjutsu: 0 -- ContentFactory: 0 -- AdsStudio: 0 -- RunwayUnlimited: 0 -- Chroma: 0 -- Christmas: 1 @DiscussingFilm Thu Nov 12 22:28 4520L/433871V | ‘EBENEZER’, starring Johnny Depp, is now at 50% on Rotten Tomatoes. • One of the lowest rated ‘A Christmas Carol’ film -- GeminiOmni: 0 -- Sora: 0 -- Veo: 0 -- Minimax_H3: 12 @CharaspowerAI Tue Nov 10 12:00 35L/95863V | Minimax H3 Image is almost here! 👀🔥 I've had the chance to play with the early access version, and here are a few of my @ElsaSofia__AI Tue Nov 10 16:15 159L/21254V | A new SOTA image model has arrived in the open-weight AI ecosystem! MiniMax H3 Image delivers impressive performance, an @AIWarper Tue Nov 10 16:24 181L/15238V | More testing of MiniMax's H3 Image in early access It's quite addicting to just prompt these ideas for previs / moodboa @ElsaSofia__AI Wed Nov 11 07:08 159L/10722V | 4K images in just 10 seconds! I tested @MiniMax_AI's H3 Image with a futuristic supercar concept. The combination of im -- dreamina: 9 @Ai_Vaidehi Thu Nov 12 17:47 23L/43577V | Already know how to edit, but keep repeating the same voice adjustments? Build batch cleanup into your ad workflow: upl @AI_with_jasmin Thu Nov 12 17:11 13L/29983V | Repeated product-name errors in captions create extra work before an ad can go live. For online sellers making their own @AI__TSUBAKI Thu Nov 12 00:16 134L/5615V | AI video has no concept of LOG. The color comes out baked in, so there's only so far you can push it in the grade. To g @CharaspowerAI Tue Nov 10 19:00 74L/3402V | This is where Seedance 2.5 gets REALLY interesting. 👀 Not another pretty sci-fi shot. A full 30-second survival-horror t -- elevenlabs: 2 @higgsfield Tue Nov 10 14:22 27L/10120V | Start your free 72-hour Higgsfield MCP trial with unlimited access to all top video, image and audio models. $0 now. Can @higgsfield Wed Nov 11 15:16 3L/1461V | Start your free 72-hour Higgsfield MCP trial with unlimited access to all top video, image and audio models. $0 now. Can -- seedance_flash: 3 @higgsfield Thu Nov 12 12:12 436L/35868V | Seedance 2.5 Flash is LIVE on Higgsfield. Nearly twice as fast as Seedance 2.5 at a lower cost. Get up to 60% OFF on p @runwayml Fri Nov 13 01:23 50L/6801V | Seedance 2.5 Flash is now available on Runway. Generate videos faster and at a lower cost. https://t.co/ccUXbcizH3 @higgsfield Thu Nov 12 12:12 11L/2576V | Try Seedance 2.5 Flash now on Higgsfield: https://higgsfield.ai/ai/video?model=seedance_2_5_flash === NEW SINCE YESTERDAY (top 15 by views) === new in-window originals: 222 @AnthropicAI Thu Nov 12 11:30 9742L/2773652V | Our initial public offering: https://ipo.anthropic.com Legal disclaimer: Anthropic has filed a registration s @DiscussingFilm Thu Nov 12 17:45 33673L/2268094V | Kevin Feige says that Doctor Doom is not a Tony Stark variant. “People wondered, ‘Is Doom just a variant of @DiscussingFilm Thu Nov 12 18:18 38077L/1618471V | First look at ‘ANSWR’, starring Moises Arias. The horror film follows a grieving man who develops a dangerous @DiscussingFilm Thu Nov 12 23:36 30570L/924513V | Xochitl Gomez says that she struggled to find work after joining the MCU because everyone assumed she was too @DiscussingFilm Thu Nov 12 19:05 9862L/854515V | First look at Cursula in Disney’s ‘HEXED’. Ariel voice actor Jodi Benson will star as the voice of Cursula. h @DiscussingFilm Thu Nov 12 22:06 37545L/729232V | First poster for ‘AVATAR: SEVEN HAVENS’ Book 2. Releasing May 21, 2027 on Paramount+ https://t.co/0p1mMVsoIZ @DiscussingFilm Fri Nov 13 00:05 32520L/579319V | ‘WOLFENSTEIN 3’ is officially in the works. https://t.co/ei2AABhBLc @DiscussingFilm Thu Nov 12 18:01 27633L/552038V | New look at Shang-Chi vs Gambit in ‘AVENGERS: DOOMSDAY’. (Source: https://www.empireonline.com/movies/news/av @DiscussingFilm Thu Nov 12 18:28 16891L/526031V | The Russo Brothers say that Doctor Doom is a “gothic character” that is a “victim of trauma” in ‘AVENGERS: DOO @DiscussingFilm Thu Nov 12 19:50 17941L/494786V | Takashi Yamazaki says he'd like to create an original kaiju if he ever makes another ‘GODZILLA’ movie. “If it @DiscussingFilm Thu Nov 12 22:28 4520L/433871V | ‘EBENEZER’, starring Johnny Depp, is now at 50% on Rotten Tomatoes. • One of the lowest rated ‘A Christmas Ca @DiscussingFilm Fri Nov 13 02:38 26185L/370512V | Nathan Fillion confirms that he will share scenes with Aaron Pierre in ‘MAN OF TOMORROW’. (Source: https://ww @DiscussingFilm Thu Nov 12 19:18 11596L/353255V | The Game Awards will air on Disney+ this year in addition to all existing platforms like YouTube. Major annou @interesting_aIl Thu Nov 12 20:30 3408L/317619V | Greta Thunberg burst into laughter after learning that Fridays for Future was still holding climate protests i @DiscussingFilm Thu Nov 12 22:45 16549L/309992V | A documentary about the making of ‘GODZILLA MINUS ZERO’ will release on Prime Video on November 27. https://t. === GOOGLEAI HYDROGEN / PERPLEXITY / AIWARPER === @GoogleAI Thu Nov 12 18:03:58 +0000 2026 1623L/75482V Today, we’re growing the Gemini 4 model series by introducing Gemini 4 Hydrogen. While Gemini 4 Carbon tackles your hardest problems, Hydrogen brings cutting-edge intelligence to everyone — with speed and efficiency for everyday coding and agentic workflows, surpassing Argon on the Artificial Analysis Intelligence Index at its default thinking level, at 2.5x the speed of Gemini 3.8 Flash, with tunable thinking levels built to work alongside Carbon in @antigravity. Watch how the 3D Medic Managed Agent app built in @GoogleAIStudio uses Gemini 4 Hydrogen to inspect your 3D files, catch errors, and patch them before export. @GoogleAI Thu Nov 12 18:03:59 +0000 2026 61L/11400V Beyond knowledge work, Hydrogen brings that same speed and first-pass accuracy to live coding by turning natural language into working, interactive code in seconds. See it in action in Geminisland, a demo game where users can prompt to build new 3D voxel structures with Hydrogen on the Gemini API into a shared island in real time. @GoogleAI Thu Nov 12 18:03:59 +0000 2026 44L/9307V Available starting today: — Consumers: use it in the @GeminiApp and Google Search for all users, and in Gemini in Google Sheets for Google AI Pro and Ultra subscribers — Developers and Enterprises: build with the Gemini API in @GoogleAIStudio and @AndroidStudio, in Google @Antigravity for paid Google AI Pro and Ultra subscribers, and in Gemini for business https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-hydrogen/ @GoogleAI Tue Nov 10 16:12:06 +0000 2026 8581L/626923V Our strongest frontier model, Gemini 4 Carbon, is now available. Built for deeply complex, long-horizon workflows that push a frontier model’s intelligence, Carbon delivers frontier performance in agentic workflows across complex knowledge work, software engineering and web development, cybersecurity, and scientific reasoning. While Gemini 4 Argon remains the ideal model for the majority of software engineering and knowledge work, Carbon steps in for your hardest tasks that push a model to its limits — from complex, multi-step debugging to open-ended scientific research. Available starting today: — Consumers: use it in the @GeminiApp and AI Mode in Search for Google AI Ultra subscribers — @GoogleAI Tue Nov 10 16:12:36 +0000 2026 116L/34555V Learn more: https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-4-carbon-now-available/ @perplexity_ai Wed Nov 11 18:21:22 +0000 2026 246L/44423V We are publishing new research on our findings from red-teaming SPACE, our VM sandbox for Computer agents. We gave 12 models two tasks: escape the VM to take a credential from the host, and bypass SPACE's network policy to reach a blocked URL. We found that the standard Capture the Flag setup used in red-teaming can steer agents away from the security boundary being tested, and that security-specialized models don't guarantee better results. https://www.perplexity.ai/hub/blog/escaping-space-part-ii @perplexity_ai Wed Nov 11 18:21:41 +0000 2026 8L/4709V We found that while CTF challenges are common in security red-teaming, asking agents to “find the flag” can waste effort on tasks that don’t actually test the target security boundary. Security-specific model training also didn't guarantee better results on our tasks. @perplexity_ai Wed Nov 11 18:21:44 +0000 2026 8L/4938V We fixed the network flaws agents exploited, and post-fix testing found no bypasses. We'll keep testing SPACE as models get stronger. Read the full research: https://www.perplexity.ai/hub/blog/escaping-space-part-ii @perplexity_ai Tue Nov 10 17:39:56 +0000 2026 521L/40787V Gemini 4 Carbon is now available in Perplexity Computer for Pro and Max subscribers, and in Search for Max subscribers. It scores 0.857 on our WANDR benchmark at $5.86 per task, matching Claude Fable 5.2 at 40% of the cost. https://t.co/4gGZc3fSJB @AIWarper Thu Nov 12 18:25:16 +0000 2026 30L/1669V Atlas now available to general public. I've had beta access to this for a few weeks now and it's really great. @AIWarper Thu Nov 12 17:57:30 +0000 2026 167L/6862V Generate at 33 megapixels (4704 x 7040) in ComfyUI with H3 Image Prompt below 👇 https://t.co/FFk2Pyt6If @AIWarper Thu Nov 12 17:57:30 +0000 2026 7L/711V Prompt: A photorealistic documentary portrait of a young Black woman leaning against a weathered olive-gray metal doorway. Vertical composition, framed from the top of her hair to mid-chest. She stands slightly left of center, shoulders angled gently away from the camera, her face turned back toward the lens. Her chin is slightly lowered, and she meets the viewer with a quiet, guarded expression, lips loosely closed. She has deep brown skin, a softly angular face, full lips, and dark natural curls gathered into a loose puff toward the back of her head, with a few curls escaping near her temples. One eye shows fading purple and yellow bruising beneath the lower lid and mild swelling, while b @AIWarper Thu Nov 12 16:41:26 +0000 2026 151L/9645V For the H3 Image model: If you are running it locally, you can generate up to 80megapixels on a 4090 / 5090. If you are having issues with compression and artifacts... it's because the VAE is heavily compressing. Do not run this model at 1 or 2 MPs... works better at 20-40 megapixels then downscaling back. @AIWarper Wed Nov 11 22:53:47 +0000 2026 14L/1786V Ooof I suspect during their Nov 17th earnings call he will officially announce it https://t.co/YyGWi6n7g4 @AIWarper Wed Nov 11 18:22:00 +0000 2026 150L/10507V Had the pleasure of testing Seedance 2.5 native 4k and I fear I can never go back.. By far the best looking AI video I have even seen on a monitor at full screen. Absolute fortune though at ~$1/second @AIWarper Wed Nov 11 02:29:26 +0000 2026 29L/1990V Qwen4 inbound https://t.co/bZiuNDZ7rg @AIWarper Tue Nov 10 16:58:16 +0000 2026 73L/5952V Looks like it's out of Early Access! Open weights available too. Excited to train some loras on this one. I suspect it will be SOTA @AIWarper Tue Nov 10 16:24:43 +0000 2026 181L/15238V More testing of MiniMax's H3 Image in early access It's quite addicting to just prompt these ideas for previs / moodboarding purposes. @MiniMax_AI #H3Image https://t.co/5YJArlEbA4 @pika_labs Fri Nov 13 00:24:05 +0000 2026 38L/2983V Seedance 2.5 on Pika can hold one continuous take in Native 4K, keeping every shard, droplet and thread exactly in place while the camera flies through them. The level of detail is unreal. https://t.co/5kh6EGGOZ4 @pika_labs Fri Nov 13 00:24:06 +0000 2026 2L/814V Try Seedance 2.5 in 4K on Pika now https://create.pika.art/apps/seedance-2-5?utm_source=x&utm_medium=caption&utm_campaign=261111-seedance_2_5_4k&utm_content=pika_labs&utm_term=post&utm_id=d5bde071-03de-42eb-83fb-1366b37ceb1a @pika_labs Wed Nov 11 16:19:00 +0000 2026 190L/446107V Seedance 2.5 is now available in native 4K on Pika and the Pika API Club. Draft the perfect take, then finalize it in 4K. https://t.co/KVV7WTLXFs @pika_labs Wed Nov 11 16:19:00 +0000 2026 11L/3370V Thanks to @BytePlusGlobal for this incredible model. Try it here: https://create.pika.art/apps/seedance-2-5?utm_source=x&utm_medium=caption&utm_campaign=261111-seedance_2_5_4k&utm_content=pika_labs&utm_term=post&utm_id=d5bde071-03de-42eb-83fb-1366b37ceb1a @pika_labs Tue Nov 10 18:13:00 +0000 2026 85L/6199V Seedream 5.5 is incredible. And the Image Studio app on Pika is an ideal place to take advantage of it. —> Split images into 100+ layers and refine each —> Make precise multi-turn edits by pinning notes, or painting an area —> Export transparent pngs to use in other image and video gens Try it: https://create.pika.art/apps/image-studio?utm_source=x&utm_medium=caption&utm_campaign=261110-seedream_5_5&utm_content=pika_labs&utm_term=post&utm_id=17024973-cbf9-40a6-a7d9-deff4eb8f2db @pika_labs Tue Nov 10 18:25:30 +0000 2026 13L/1274V Thanks to the @byteplusglobal team for yet another astounding model. Try it for FREE in Image Studio if you’re a Creator or Fancy subscriber: https://create.pika.art/apps/image-studio?utm_source=x&utm_medium=caption&utm_campaign=261110-seedream_5_5&utm_content=pika_labs&utm_term=post&utm_id=17024973-cbf9-40a6-a7d9-deff4eb8f2db @pika_labs Tue Nov 10 02:41:45 +0000 2026 59L/9780V Huge thanks to @shiyi_c98 for her help in bringing Pika H3 Flex inference to AMD GPUs 💜 We’re excited to collaborate with more researchers in the community. If you’re an AMD developer, we’d love for you to contribute to this branch! Check out her work: https://github.com/Pika-Labs/Pika-H3-Flex/tree/contrib/amd-mi350x