正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

Same perfume concept. Four video models. The difference becomes obvious once the scene starts changing. Vivix W1 keeps the subject and product feeling coherent as the studio transitions into the storm. For me, that matters more than how dramatic the storm looks. An eight-second version costs $0.024 to generate.... Waiting for it takes roughly eight seconds. This comparison only shows a finished clip. W1 is built for real-time interactive video, not just one-shot generation. Vivix.AI already offers that real-time capability through the API. At $0.003/sec, that also makes experimenting with interactive video workflows a lot more accessible. Letting a viewer influence how a scene unfolds would mean paying for generation throughout the interaction. Worth watching the comparison closely. The model behind the studio-to-storm transition 👇show more

Liam | AI Tools & News

50,945 subscribers

42,351 次观看 • 5 天前 •via X (Twitter)

教育科学技术

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

this changes what AI video can even be $0.003 per second of AI video. that's not a typo. for the price of ONE Seedance 2.5 video, you can generate almost 80 W1 videos of the same length. and here's the part that makes it insane: it's not slower and it's not lower quality. an 8 second video generates in about 8 seconds, at quality comparable to the top models. cheaper, fast, and good. usually you pick two. Vivix.AI just brought all three together. but the price isn't even the real story. W1 is a real-time interactive video model. video that generates as an experience unfolds, so you can interact with what happens next. real-time means it keeps generating the whole time someone's using it. longer sessions, more users, more video, more cost. that only works if every second is cheap. $0.003/sec is what finally makes interactive video practical to actually run. i want to make a scene where even i don't know how it ends. honestly, giving up that control sounds fun. (real-time is live via the API now, browser tools coming soon) save this.

this changes what AI video can even be $0.003 per second of AI video. that's not a typo. for the price of ONE Seedance 2.5 video, you can generate almost 80 W1 videos of the same length. and here's the part that makes it insane: it's not slower and it's not lower quality. an 8 second video generates in about 8 seconds, at quality comparable to the top models. cheaper, fast, and good. usually you pick two. Vivix.AI just brought all three together. but the price isn't even the real story. W1 is a real-time interactive video model. video that generates as an experience unfolds, so you can interact with what happens next. real-time means it keeps generating the whole time someone's using it. longer sessions, more users, more video, more cost. that only works if every second is cheap. $0.003/sec is what finally makes interactive video practical to actually run. i want to make a scene where even i don't know how it ends. honestly, giving up that control sounds fun. (real-time is live via the API now, browser tools coming soon) save this.

Khairallah AL-Awady

29,796 次观看 • 3 天前

I spent the morning running cost projections for live interactive video, and the math on conventional models is completely broken for production. That is why Vivix.AI's W1 model caught my attention. At $0.003/sec, I can trigger nearly 77 equal-length W1 generations for the API cost of a single Seedance 2.5 call. When you build applications where video generates non-stop as users interact, unit economics dictate whether your project survives. This affordability is what makes W1’s identity as a real-time interactive video model practically usable. API access is active today, though browser playground tools are currently offline generation only while real-time updates roll out. Looking at their 4-way comparison with Seedance 2.5, Kling 3.0, and H3 Max fal featuring the slow-motion sneeze window shatter, W1 holds its visual fidelity remarkably well. You still see minor physical artifacts in high-density glass debris while full-scale model training progresses, but solving the real-time inference cost curve is the exact infrastructure breakthrough I was waiting for. The numbers that put W1 on my production shortlist 👇

I spent the morning running cost projections for live interactive video, and the math on conventional models is completely broken for production. That is why Vivix.AI's W1 model caught my attention. At $0.003/sec, I can trigger nearly 77 equal-length W1 generations for the API cost of a single Seedance 2.5 call. When you build applications where video generates non-stop as users interact, unit economics dictate whether your project survives. This affordability is what makes W1’s identity as a real-time interactive video model practically usable. API access is active today, though browser playground tools are currently offline generation only while real-time updates roll out. Looking at their 4-way comparison with Seedance 2.5, Kling 3.0, and H3 Max fal featuring the slow-motion sneeze window shatter, W1 holds its visual fidelity remarkably well. You still see minor physical artifacts in high-density glass debris while full-scale model training progresses, but solving the real-time inference cost curve is the exact infrastructure breakthrough I was waiting for. The numbers that put W1 on my production shortlist 👇

Eleftheria Batsou

40,025 次观看 • 3 天前

Letting Michael keep improvising sounds expensive. At $0.003 per generated second, Vivix W1 at least makes the video part more manageable. That's why his switch from “pick two” to “pick three” lands: a real-time interactive video model needs to afford the next unscripted moment while generating quickly and maintaining comparable visual quality. Watching Dwight spray him, I'm less confident the office has budgeted for everything else.

Letting Michael keep improvising sounds expensive. At $0.003 per generated second, Vivix W1 at least makes the video part more manageable. That's why his switch from “pick two” to “pick three” lands: a real-time interactive video model needs to afford the next unscripted moment while generating quickly and maintaining comparable visual quality. Watching Dwight spray him, I'm less confident the office has budgeted for everything else.

Alamin

41,013 次观看 • 5 天前

one of the easiest ways to make your AI videos look 10x better is by just increasing the saturation a bit > left is the original output > right is saturation + contrast increased for some reason AI video models love making everything feel cold and blue, which makes the video look obviously AI very quickly a bit of saturation fixes a lot of that then pair it with Topaz and a bitrate cleaner and the video starts looking like it was shot on a real iphone cam obviously you still need a good system behind the actual video, but this takes a good generation to a completely different level

one of the easiest ways to make your AI videos look 10x better is by just increasing the saturation a bit > left is the original output > right is saturation + contrast increased for some reason AI video models love making everything feel cold and blue, which makes the video look obviously AI very quickly a bit of saturation fixes a lot of that then pair it with Topaz and a bitrate cleaner and the video starts looking like it was shot on a real iphone cam obviously you still need a good system behind the actual video, but this takes a good generation to a completely different level

Miko

17,295 次观看 • 1 个月前

Sora 2, “ family guy dark humor” My personal thoughts: This is clearly the best AI video model for replicating animated shows. This truly gives credence to the fact that within the decade AI generated short films/shows is looking like a real possibility. My important caveats. Model trained only on video can get visually indistinguishable for most viewers most of the time. It can match texture, lighting, motion, and camera language so well that only careful inspection gives it away. However For sustained quality at the level of a full scene or an episode, video alone hits a ceiling IMO. Pure video likelihood drives the model toward what is frequent, not toward the rare timing and payoff choices that make the best jokes land. It has weak grasp of long arc causality, character memory, and joke structure. It also does not see intent, off screen context, or prosody unless you give it those signals. So you get something that looks right but drifts on beats that matter. I don’t know what the potential solution would be other than to have an AGI just animate the show for me. Any others ? Credit for the sora clip: figure

Sora 2, “ family guy dark humor” My personal thoughts: This is clearly the best AI video model for replicating animated shows. This truly gives credence to the fact that within the decade AI generated short films/shows is looking like a real possibility. My important caveats. Model trained only on video can get visually indistinguishable for most viewers most of the time. It can match texture, lighting, motion, and camera language so well that only careful inspection gives it away. However For sustained quality at the level of a full scene or an episode, video alone hits a ceiling IMO. Pure video likelihood drives the model toward what is frequent, not toward the rare timing and payoff choices that make the best jokes land. It has weak grasp of long arc causality, character memory, and joke structure. It also does not see intent, off screen context, or prosody unless you give it those signals. So you get something that looks right but drifts on beats that matter. I don’t know what the potential solution would be other than to have an AGI just animate the show for me. Any others ? Credit for the sora clip: figure

Chris

150,286 次观看 • 1 年前

I made a second version of the transformation video. This time I moved the character into a real environment instead of using a plain white background. I also changed the opening and showed the final transformation result first. The overall flow feels much stronger now. Full workflow + prompts: The interesting part is that this type of video is much easier to make than it looks. You only need a base image, then generate the next keyframe through simple edits or outfit changes on the canvas. The entire transformation sequence is built from those keyframes.

I made a second version of the transformation video. This time I moved the character into a real environment instead of using a plain white background. I also changed the opening and showed the final transformation result first. The overall flow feels much stronger now. Full workflow + prompts: The interesting part is that this type of video is much easier to make than it looks. You only need a base image, then generate the next keyframe through simple edits or outfit changes on the canvas. The entire transformation sequence is built from those keyframes.

underwood

315,443 次观看 • 4 个月前

There’s a weird paradox in AI video right now: The more ambitious the shot, the less you want the model improvising everything. For a simple clip, “describe → generate” can be enough. But for a complex scene, it helps to define the world first. That’s the idea behind the GPT-6 Astra workflow I’ve been testing: build the 3D scene, establish spatial relationships, and define the camera and character paths in the Viewport. Then, manually bring everything into CapCut PC. Seedance 2.5 handles the video rendering, while CapCut PC takes the result through post-production. Spatial control before generation. Creative control after it. #CapCutPC #GPT6Astra #AIVideo #Seedance25

There’s a weird paradox in AI video right now: The more ambitious the shot, the less you want the model improvising everything. For a simple clip, “describe → generate” can be enough. But for a complex scene, it helps to define the world first. That’s the idea behind the GPT-6 Astra workflow I’ve been testing: build the 3D scene, establish spatial relationships, and define the camera and character paths in the Viewport. Then, manually bring everything into CapCut PC. Seedance 2.5 handles the video rendering, while CapCut PC takes the result through post-production. Spatial control before generation. Creative control after it. #CapCutPC #GPT6Astra #AIVideo #Seedance25

Fathela Esq

43,328 次观看 • 1 个月前

BREAKING: Hurricane Hunters just measured an almost unbelievable 889 mb pressure inside Hurricane Polo. If that reading is confirmed, Polo would become the second strongest hurricane ever recorded in the eastern Pacific, behind only Hurricane Patricia in 2015. What makes this even more remarkable is how quickly Polo got here. The storm has intensified at an extraordinary rate, being only a tropical storm 24 hours ago. An 889 mb pressure is an exceptionally rare number in the eastern Pacific, putting Polo into truly historic territory. We are watching one of the strongest Pacific hurricanes ever observed unfold in real time.

BREAKING: Hurricane Hunters just measured an almost unbelievable 889 mb pressure inside Hurricane Polo. If that reading is confirmed, Polo would become the second strongest hurricane ever recorded in the eastern Pacific, behind only Hurricane Patricia in 2015. What makes this even more remarkable is how quickly Polo got here. The storm has intensified at an extraordinary rate, being only a tropical storm 24 hours ago. An 889 mb pressure is an exceptionally rare number in the eastern Pacific, putting Polo into truly historic territory. We are watching one of the strongest Pacific hurricanes ever observed unfold in real time.

Max Velocity

4,343,981 次观看 • 22 天前

We just launched Claude Sonnet 4.5 and “Imagine with Claude”, a 5-day preview that creates dynamic experiences for you in real-time (and the first product I’ve worked on at Anthropic!). It pioneers the concept of “model-as-backend”, using a model to not only generate interfaces on the fly but also power all the functionality behind it – all made possible by our newest model, Claude Sonnet 4.5. Here’s a "choose your own adventure" version of my founder journey, as retold by Imagine with Claude:

We just launched Claude Sonnet 4.5 and “Imagine with Claude”, a 5-day preview that creates dynamic experiences for you in real-time (and the first product I’ve worked on at Anthropic!). It pioneers the concept of “model-as-backend”, using a model to not only generate interfaces on the fly but also power all the functionality behind it – all made possible by our newest model, Claude Sonnet 4.5. Here’s a "choose your own adventure" version of my founder journey, as retold by Imagine with Claude:

Sean Strong

115,727 次观看 • 1 年前

A CHINESE STUDENT REMAKES OLD TOM AND JERRY SCENES WITH REAL ANIMALS AND MADE $13,286 LAST MONTH > 263,000 likes on 8,100 comments. Cartoon on top, live action below, Tom in a straw hat holds out a ham, a lion bites it from behind a blue curtain, and a real cat and lion copy every beat. The cartoon is the shape and the real cat is the only thing that changed: -> Tom's straw hat with the green band reads in under a second, so the brain loads the whole scene before it plays. -> The ham gets bitten down to the bone in both rows and the real cat grabs the same rifle Tom does, the detail people pause to compare. -> The lion hides behind the same blue curtain in both versions, so the swap reads as a remake of the exact scene. -> The shape is a scene millions already watched, the subject is a real cat nobody has seen in it, and that pairing is what the sheet hunts for. -> The comments fill with people naming the episode while the clip scores against the channel median and anything past 35 gets the next scene that week. A cartoon people grew up on is a format that already proved itself. Rebuild it real and the old scene pays again. It works on any cartoon scene people know frame by frame, a Road Runner chase, a Scooby-Doo hallway, a Looney Tunes anvil, a Pink Panther tiptoe. Claude picks the scene and writes the 6 shots, CapCut builds the real version, Make posts it and reports back for $49 a month. The real cat got the likes, and the cartoon everyone grew up on is what got them watching. The sheet that picks every scene is below 👇

A CHINESE STUDENT REMAKES OLD TOM AND JERRY SCENES WITH REAL ANIMALS AND MADE $13,286 LAST MONTH > 263,000 likes on 8,100 comments. Cartoon on top, live action below, Tom in a straw hat holds out a ham, a lion bites it from behind a blue curtain, and a real cat and lion copy every beat. The cartoon is the shape and the real cat is the only thing that changed: -> Tom's straw hat with the green band reads in under a second, so the brain loads the whole scene before it plays. -> The ham gets bitten down to the bone in both rows and the real cat grabs the same rifle Tom does, the detail people pause to compare. -> The lion hides behind the same blue curtain in both versions, so the swap reads as a remake of the exact scene. -> The shape is a scene millions already watched, the subject is a real cat nobody has seen in it, and that pairing is what the sheet hunts for. -> The comments fill with people naming the episode while the clip scores against the channel median and anything past 35 gets the next scene that week. A cartoon people grew up on is a format that already proved itself. Rebuild it real and the old scene pays again. It works on any cartoon scene people know frame by frame, a Road Runner chase, a Scooby-Doo hallway, a Looney Tunes anvil, a Pink Panther tiptoe. Claude picks the scene and writes the 6 shots, CapCut builds the real version, Make posts it and reports back for $49 a month. The real cat got the likes, and the cartoon everyone grew up on is what got them watching. The sheet that picks every scene is below 👇

Elvar

1,567,224 次观看 • 13 天前

$LTX 2.5 JUST WENT FASTER THAN REAL TIME LTX (LTX.io) is already one of the world's most widely used open video models. We just made its latest release, LTX 2.5, fast enough for rapid iteration and affordable enough for production at scale. It’s live on TensorScaleAI now, generating a 5-sec 1080p video in just 3.12s at $0.04 per second. That's faster than real time, at a fraction of the cost! For creative teams, that means less waiting time, more ideas explored, and lower production costs. For robotics and physical AI teams, that means more data generated at scale, more scenarios tested, and shorter development and deployment cycles. Request early access today and we’ll get you onboarded! (link:$

LTX 2.5 JUST WENT FASTER THAN REAL TIME LTX (LTX.io) is already one of the world's most widely used open video models. We just made its latest release, LTX 2.5, fast enough for rapid iteration and affordable enough for production at scale. It’s live on TensorScaleAI now, generating a 5-sec 1080p video in just 3.12s at $0.04 per second. That's faster than real time, at a fraction of the cost! For creative teams, that means less waiting time, more ideas explored, and lower production costs. For robotics and physical AI teams, that means more data generated at scale, more scenarios tested, and shorter development and deployment cycles. Request early access today and we’ll get you onboarded! (link:

Yao Zhang

16,733 次观看 • 1 个月前

I think people are still underestimating how important the 3D layer could become in AI video. The problem with complex AI shots has never just been image quality. It’s spatial consistency, camera direction and knowing where everything is supposed to be. GPT-6 Astra can map the scene and keep camera + character paths within the Viewport, giving the shot a spatial structure before generation. I then manually move that scene into CapCut PC and use Seedance 2.5 to render it. From there, CapCut PC becomes the actual editing environment. That feels less like “AI generated a video” and more like a production pipeline worth experimenting with. #CapCutPC #GPT6Astra #AIVideo #Seedance25

I think people are still underestimating how important the 3D layer could become in AI video. The problem with complex AI shots has never just been image quality. It’s spatial consistency, camera direction and knowing where everything is supposed to be. GPT-6 Astra can map the scene and keep camera + character paths within the Viewport, giving the shot a spatial structure before generation. I then manually move that scene into CapCut PC and use Seedance 2.5 to render it. From there, CapCut PC becomes the actual editing environment. That feels less like “AI generated a video” and more like a production pipeline worth experimenting with. #CapCutPC #GPT6Astra #AIVideo #Seedance25

Manish Kumar Shah

30,785 次观看 • 1 个月前

Seedance 2.0 + Claude Code is f*cking insane 🤯 I built a Claude skill that creates UGC ads on demand. One product + one prompt = the AI creator, the script, the scene-by-scene shot list, and the finished video. All inside Claude Code. Perfect for DTC brands and agencies who can't afford to keep paying $500-$1,500 per UGC video and waiting 2 weeks for revisions. This skill eliminates the entire loop: → Tell Claude the product, ad angle, and length → Skill writes the GPT Image 2.0 prompt to generate the AI creator from scratch → Skill writes every scene prompt, dialogue line, and delivery direction → Pipes it into Seedance 2.0 with character + product + voice locked → Speed up + caption in CapCut → Ship the ad in 20 minutes No more paying $11 per video on Arcads. No more 2-week revision cycles. No more PR boxes to creators who ghost you. What you get: → Perfect character consistency across every scene → Voice consistency that holds clip-to-clip → Real product fidelity using your actual product photo as a reference → Multi-scene day-in-life, testimonial, and action-shot formats out of the box Built 100% with a Claude skill + Seedance 2.0. I recorded a full step-by-step tutorial showing the exact workflow so you can build these AI UGC ads yourself. Want the full breakdown? > Like this post > Comment "UGC" And I'll send it over (must be following so I can DM)

Seedance 2.0 + Claude Code is f*cking insane 🤯 I built a Claude skill that creates UGC ads on demand. One product + one prompt = the AI creator, the script, the scene-by-scene shot list, and the finished video. All inside Claude Code. Perfect for DTC brands and agencies who can't afford to keep paying $500-$1,500 per UGC video and waiting 2 weeks for revisions. This skill eliminates the entire loop: → Tell Claude the product, ad angle, and length → Skill writes the GPT Image 2.0 prompt to generate the AI creator from scratch → Skill writes every scene prompt, dialogue line, and delivery direction → Pipes it into Seedance 2.0 with character + product + voice locked → Speed up + caption in CapCut → Ship the ad in 20 minutes No more paying $11 per video on Arcads. No more 2-week revision cycles. No more PR boxes to creators who ghost you. What you get: → Perfect character consistency across every scene → Voice consistency that holds clip-to-clip → Real product fidelity using your actual product photo as a reference → Multi-scene day-in-life, testimonial, and action-shot formats out of the box Built 100% with a Claude skill + Seedance 2.0. I recorded a full step-by-step tutorial showing the exact workflow so you can build these AI UGC ads yourself. Want the full breakdown? > Like this post > Comment "UGC" And I'll send it over (must be following so I can DM)

Mike Futia

41,849 次观看 • 5 个月前

THIS GUY VIBE CODED A BROWSER TOOL WITH CLAUDE CODE THAT DRAWS PROPERTY LINES ONTO DRONE FOOTAGE AND KEEPS THEM LOCKED TO THE GROUND AS THE DRONE MOVES real estate agents kept asking for those property boundary overlays on drone videos so buyers can see exactly where a lot ends. the only real option was after effects, which is a massive pro tool for one small repeatable job so he built a browser tool that does just that one thing > drop in a drone clip and draw the parcel outline once on the first frame > gpu motion tracking keeps that boundary stuck to the ground as the drone flies and the perspective shifts > if the line drifts anywhere, you drop a keyframe to correct it and it re-tracks the segments around it > style the line however you want, then export the finished video > free to use, you dont even need an account until you export he built one tool that solves one annoying job perfectly, for people who were never going to learn after effects anyway

THIS GUY VIBE CODED A BROWSER TOOL WITH CLAUDE CODE THAT DRAWS PROPERTY LINES ONTO DRONE FOOTAGE AND KEEPS THEM LOCKED TO THE GROUND AS THE DRONE MOVES real estate agents kept asking for those property boundary overlays on drone videos so buyers can see exactly where a lot ends. the only real option was after effects, which is a massive pro tool for one small repeatable job so he built a browser tool that does just that one thing > drop in a drone clip and draw the parcel outline once on the first frame > gpu motion tracking keeps that boundary stuck to the ground as the drone flies and the perspective shifts > if the line drifts anywhere, you drop a keyframe to correct it and it re-tracks the segments around it > style the line however you want, then export the finished video > free to use, you dont even need an account until you export he built one tool that solves one annoying job perfectly, for people who were never going to learn after effects anyway

Om Patel

404,227 次观看 • 2 个月前

$Fable 5 built this real-time ocean scene directly in the browser using Three.js, WebGPU and TSL, with no engine and no baked assets. The interesting part is not just that the water looks good. The whole system is being simulated: wave spectrum, GPU FFT, foam generation and refracted-ray caustics, all running live. This is exactly where frontier models start becoming genuinely useful for technical graphics work. Not by replacing years of rendering research, but by compressing the path from complex math to a working interactive result. A small, focused scene, but technically very serious. Cc: Adem Vessell$

Fable 5 built this real-time ocean scene directly in the browser using Three.js, WebGPU and TSL, with no engine and no baked assets. The interesting part is not just that the water looks good. The whole system is being simulated: wave spectrum, GPU FFT, foam generation and refracted-ray caustics, all running live. This is exactly where frontier models start becoming genuinely useful for technical graphics work. Not by replacing years of rendering research, but by compressing the path from complex math to a working interactive result. A small, focused scene, but technically very serious. Cc: Adem Vessell

Token Gremlin

21,731 次观看 • 3 个月前

1/ We've all been aware of the hype surrounding Dreamina Seedance 2.0, and I finally got early access to the tool. And wow, I'm blown away. This is by far the best. It is starting to make AI video feel less like "generate a clip" and more like "direct a scene." What stood out to me is the level of control: camera motion, pacing, visual consistency, and the ability to build from multiple references inside one workflow. Some of the prompt directions that feel especially strong: - a busy modern city square during daytime. Suddenly, time freezes completely - a single continuous camera movement through a natural landscape that transitions through all four seasons in one shot - an underwater bioluminescent city waking up at dawn The big shift is this: One Prompt, Viral Remade. Edit Videos as Easy as Editing Photos. Dreamina Seedance 2.0 feels like a real step toward AI-native directing rather than just AI generation. Here are some examples 🧵:

1/ We've all been aware of the hype surrounding Dreamina Seedance 2.0, and I finally got early access to the tool. And wow, I'm blown away. This is by far the best. It is starting to make AI video feel less like "generate a clip" and more like "direct a scene." What stood out to me is the level of control: camera motion, pacing, visual consistency, and the ability to build from multiple references inside one workflow. Some of the prompt directions that feel especially strong: - a busy modern city square during daytime. Suddenly, time freezes completely - a single continuous camera movement through a natural landscape that transitions through all four seasons in one shot - an underwater bioluminescent city waking up at dawn The big shift is this: One Prompt, Viral Remade. Edit Videos as Easy as Editing Photos. Dreamina Seedance 2.0 feels like a real step toward AI-native directing rather than just AI generation. Here are some examples 🧵:

Chubby♨️

76,702 次观看 • 6 个月前

The AI video race is getting crowded. 👀 Seedance, Kling, H3, and now Vidu's Q-series finally gets an update after 8 months. Vidu Q4 Preview just dropped, and there are a few things that caught my attention. The standout for me: consistency. Q4 Preview supports up to 15 reference images to keep characters, products, and visuals consistent. That’s a lot more control when building a scene. It also supports up to 3 reference audio tracks, with generations of 4K output. Starting at $0.014/second, much cheaper than seedance That’s where the economics get interesting. The AI video race isn’t just about who can generate the best-looking clip anymore. It’s also about how easily creators can experiment, iterate, and turn more ideas into actual videos. Vidu Q4 Preview is definitely one to watch. The AI video race is getting interesting. 👀 With more capable models and lower generation costs, creators have more room to experiment. Which AI video model are you betting on right now? #ViduQ4Preview #ViduAI

The AI video race is getting crowded. 👀 Seedance, Kling, H3, and now Vidu's Q-series finally gets an update after 8 months. Vidu Q4 Preview just dropped, and there are a few things that caught my attention. The standout for me: consistency. Q4 Preview supports up to 15 reference images to keep characters, products, and visuals consistent. That’s a lot more control when building a scene. It also supports up to 3 reference audio tracks, with generations of 4K output. Starting at $0.014/second, much cheaper than seedance That’s where the economics get interesting. The AI video race isn’t just about who can generate the best-looking clip anymore. It’s also about how easily creators can experiment, iterate, and turn more ideas into actual videos. Vidu Q4 Preview is definitely one to watch. The AI video race is getting interesting. 👀 With more capable models and lower generation costs, creators have more room to experiment. Which AI video model are you betting on right now? #ViduQ4Preview #ViduAI

RAVI KUMAR SAHU

135,615 次观看 • 7 天前

Wan 3.0 is live on Magnific. One generation now takes 20 references at once: 10 images, 5 audio, 5 videos. Not one style ref. Twenty. What that actually changes: Length : 2 to 30 seconds in a single pass. No stitching, no extending clip by clip and praying the character survives the cut. Native audio : voice, ambience and SFX generated with the video, not laid on top afterwards. Multiple ratios available , and adaptive the model composes for the frame instead of cropping into it. 10 image refs : a character, a product, a location and a lighting look can all hold in the same shot. ⚡ 40% off until September 30. More details in comments 👇 What's the first thing you'd throw 20 references at?

Wan 3.0 is live on Magnific. One generation now takes 20 references at once: 10 images, 5 audio, 5 videos. Not one style ref. Twenty. What that actually changes: Length : 2 to 30 seconds in a single pass. No stitching, no extending clip by clip and praying the character survives the cut. Native audio : voice, ambience and SFX generated with the video, not laid on top afterwards. Multiple ratios available , and adaptive the model composes for the frame instead of cropping into it. 10 image refs : a character, a product, a location and a lighting look can all hold in the same shot. ⚡ 40% off until September 30. More details in comments 👇 What's the first thing you'd throw 20 references at?

LudovicCreator

24,012 次观看 • 1 个月前

Is 3D becoming the missing middle layer between AI imagination and controllable video? Text can describe what a scene should look like. Video models can render it. But something still has to define the spatial relationships in between. That’s where GPT-6 Astra becomes interesting: it can understand the 3D environment and establish camera and character movement inside the Viewport. The resulting scene is manually brought into CapCut PC, where Seedance 2.5 handles the video generation. Then the generated footage can continue into CapCut PC for editing. It’s not really “3D replacing AI video.” It’s more like giving AI video a spatial blueprint to follow. That feels like a direction worth experimenting with in CapCut PC. #CapCutPC #GPT6Astra #AIVideo #Seedance25

Is 3D becoming the missing middle layer between AI imagination and controllable video? Text can describe what a scene should look like. Video models can render it. But something still has to define the spatial relationships in between. That’s where GPT-6 Astra becomes interesting: it can understand the 3D environment and establish camera and character movement inside the Viewport. The resulting scene is manually brought into CapCut PC, where Seedance 2.5 handles the video generation. Then the generated footage can continue into CapCut PC for editing. It’s not really “3D replacing AI video.” It’s more like giving AI video a spatial blueprint to follow. That feels like a direction worth experimenting with in CapCut PC. #CapCutPC #GPT6Astra #AIVideo #Seedance25

Onil Coder

12,000 次观看 • 1 个月前