I Tried Every New AI Video Generator And THIS Is The One You Should Use

I ran Kling 3.0, SeaArt 2.5, Gen-1 Omni Flash, and Flex 3 through the same seven tests to see which AI video generator is actually the best and worth your money. Every model is using the exact same prompt and the reference image and the closest possible settings. And the results were way more different than I ever anticipated. So, make sure you stick around because by the end of this video, I'm going to tell you exactly which model is best for what and which one is actually worth your hard-earned money. Now, for this, we are going to be using Higgs Field to generate all of our different videos because Higgs Field has all these different models for us to choose from. Okay, so we're going to be testing these four different models on a few different parameters. And that's going to include realism, character consistency, UGC, prompt adherence, motion quality, lip sync, cinematic quality, and of course, credits and the cost to see if any one of these models are actually worth your money or not. All right, so with that, let's just go ahead and get started with our first test, which is going to be the realism test. Now, for the realism test, ahead and generated a reference image of a barista simply just handing out a coffee. And as for the prompt for our video, we have the barista slides the coffee towards the customer, wipes a little spilled coffee from the counter, and turns around to start the next order. People in the background continue talking and drinking their coffee naturally. The camera stays still like someone casually recorded it from across the counter. So, let's go ahead and check out our first generation. We're going to be starting off with Flex 3. All right, so overall, the entire feel of the video is pretty good. All of the letters, all of the wording, the pricing even is good on the little signage up here. The little coffee back here, these coffee dispensers look a little bit off. They don't look like they have the little part where the coffee actually comes down from. But besides that, I find it a little bit strange the way that she just stares at the coffee and then at the end, when she goes to grab the cup, it kind of moves a little bit awkwardly and then it turns into a napkin. So, I think that's the only thing that really stood out to me. Besides that, everything else looks pretty good. The people in the background look good. The person in the back with the hat, he seems to be drinking coffee without a coffee cup. But, besides that, everything else looks pretty good. So, small details, nothing too crazy that really threw me off. So, for realism for Flux, let's go ahead and give it a 7 out of 10. I thought it was pretty good. All right, so now let's check out Google Omniflux. Okay, so this one so far was actually not the best. The first thing that really stood out was the cup jumped places. That was very noticeable. Then, everything kind of just felt like paper, like let's say the sign in the back, it just had this weird little wave that went through it. And then, the same thing happened to the window itself. And everything kind of just pulsed or pulsated for a bit. Also, when she turns around and goes to make the coffee and grabs the little stainless steel cup in the back, it's kind of awkward on that part. She didn't really grab it the right way. So, out of 10, Gemini Omniflux, I do want to say is probably for this one, it didn't look realistic at all, especially with that weird little pulse. I'm going to give it a 3 out of 10. All right, so up next, we're going to try out Kling 3.0. Everything sounded pretty well. I like the ambient noise. Everything about the scene in the background was pretty good, except for the guy. He seemed to kind of just be talking to himself. With the chair that's in front of him, looks like it has a coat on it, but it's a little bit weird. It's a little bit off. It doesn't look right. But, then the biggest and most obvious thing that was off with this entire video is just the way that she's presenting or passing over the cup. She turns it around and it magically ends up having two handles. It kind of moves a little bit by itself. And then the coffee spill spilled on the table, but none of it fell on the actual cup or the plate. So, while the quality looks pretty good, the realism of it wasn't really there. So, Kling, unfortunately, I am going to give it a lower score than I anticipated to give it. I'm going to give it a three out of 10 as well. All right, so now let's try out Sea Dance 2.5. Okay, honestly, I found absolutely nothing wrong with this. Everything came out perfect. This looks like it could be legitimately real footage. One thing that I really liked, that I felt like was the icing on the cake, is that if you see on her nail, she has a little black dot on it, kind of like it was dirty. And that just adds another level of realism to me for that. Besides that, the ambient noise of people in the background hanging out, everything was pretty spot-on. Honestly, I think there wasn't anything bad with this other than the fact that it generated in landscape instead of portrait. But I did have the orientation set to automatic, so some of these did flip over, but that's not the model's fault. Honestly, I think this is flawless. This one's perfect to me. I'm going to give Sea Dance for this one a 10 out of 10. All right, next we're going with character consistency. And for this, I generated an image of a woman in a record store wearing sunglasses. So, for the character consistency, the way that we want to keep it consistent is by keeping an eye on the prompt and what they're actually doing. So, as for the prompt, we have she walks over to the shelf and looks through a few records. She finds one she likes, takes off her green sunglasses, and smiles while looking at it. Keep her face, hair, clothes, earrings, and flower patch looking the same throughout the video. So, this is our reference image, and we're going to be starting in this one again with Flux 3. So, everything from this video was pretty good. The only thing is when she actually went to take off her glasses. So, when she takes off her glasses, she grows a third arm, and that right there kind of just threw me off for a second. And then the arm just disappears again. It retracts. Everything else does stay the same. The glasses look pretty good throughout the entire video. When she puts them down in front of the record, the feet or the legs on it, they do look like they're upside down. So, that part is not consistent. Herself though, she is consistent all the way up until she grows that third arm. And one more thing that actually got me was if we stop at about 3 seconds, we have a creepy lady on the corner who just is staring weirdly and then disappears. As far as the character consistency, again, everything was good up until that third arm came up. And the glasses also kind of threw me off when she flipped them over. So, I am going to give it a two out of 10. If it was just the glasses, it wouldn't have been that bad, but that third arm, that just kills everything. All right, next up, let's try out Google Omni Flash. >> [music] [music] >> Okay, so Google Omni Flash actually did really, really good. When it changed angles, it was still the same person. Everything looks exactly the same. She has her nose ring on You know what? She didn't have a nose ring in the first two scenes, now that I'm looking back at it. And she then has a nose ring in the third scene when she takes off her glasses. So, that's not a huge deal. That's probably something that a lot of people might not pick up on, but I did notice it. So, that is going to deduct points off of it. But, besides that, everything else was really good. Hair stayed the same, body stayed the same, same jacket, same flower, same glasses. It's just the nose ring that was off. Even the background and everything is looking to be the exact same records as well. So, Omni-Flash is definitely getting an upgraded score from the last one, and we are going to give them an eight out of 10. It was really good. All right, next up, let's go with Kling 3. >> [sighs] >> All right, so Kling did a really good job. The only thing that was a little bit strange to me was how she was looking at the record. She looked at it, flipped it maybe twice, and then picked it up and looked at it. It's a little bit unnatural, but that's not the point of this particular test. The point of this test is the character consistency, and everything did stay consistent throughout the entire process. Kling actually did really well, and I'm going to give it a nine out of 10. The way that she looked at the record kind of threw me off a little bit, but again, we're not scoring it based on that. The character consistency was pretty spot-on. All right, so now let's check out C Dance 2.5. So, C Dance 2.5 had a very consistent character throughout the entire video. Everything looked very realistic. Everything looked very relaxed, very chill. It was very smooth. The way that she looked through the records was a little bit off, maybe, but it wasn't too bad, to be honest with you. And then when she pulls it out, she pulls up her glasses. Everything stays the same. There's no added nose ring. The rings that are on her hand all are there, and the necklace that she started off with at the beginning of the frame is there as well. So, overall, C Dance did an amazing job, and I'm actually going to give it a 10 out of 10. This one was flawless. All right, so for the next test, we are going to be testing out something that we can all use in any one of our stores or for businesses, and that's AI UGC. Now, I'm going to be doing specifically an e-commerce type UGC video. So, I did have to combine two different images in Higgs Field. I had to give the reference of the product and the prompt to create the woman holding the product. So, there was a little bit of an extra added step, but no big deal. It just took a couple of extra more minutes. So, for our AI UGC, the prompt that we're using is going to be she holds the bottle closer to the camera, turns it slightly to show the packaging, and then brings it back beside her face. She says, "I've been using this for about a week now, and honestly, I didn't expect it to like it this much. Make it feel like a casual TikTok someone is recording at home." Now, let's start with Flux for this one. >> I've been using this for about a week now, and honestly, I didn't expect to like it this much. >> Okay, so overall, it's not that bad the way that she shows off the product. The product itself is consistent throughout the entire video. When she moves it back towards her face, the name of the brand and all of the text on the product itself do start to blur. I did choose a Neutrogena product for this cuz I did want to see how the logo and everything stayed and how it was consistent. Another thing that's throwing me off is the way that she speaks. As she's talking, she kind of has these weird little breaks. She talks a little bit slow, and the focus on her eyes are supposed to be on the phone as if she's doing a UGC video, and they are on there, but something about it just looks unnatural. It's not as she's focusing directly on the phone where she should be. This could probably work with a little bit of editing, a little bit of speeding up. It could probably make it a little bit better, but right off the bat, honestly, I'm not too impressed. I've seen way better results. One specifically with a platform that is actually dedicated for UGC, which I'm actually going to show you in just a little bit, so make sure you stick around for that. But, as far as Flux, I wasn't too impressed. It wasn't the worst. With a little bit of editing, again, it could get fixed. So, let's go ahead and give it a five out of 10. All right, now we're going to go ahead and check it out on Cling 3. >> I've been using this for about a week now, and honestly, I didn't expect to like it this much. >> All right, so Cling really disappointed me right off the bat. I really did not like at all the way that the bottle looks. It's already being distorted. Even when she holds it close to the camera, it looks a little bit off. Another thing is the way that she isn't talking at the beginning, it's very hesitant, it's very awkward, it's not working at all for me. >> And honestly, I didn't expect to like it this much. >> And then when she does start to talk, her voice isn't fully synced to her mouth. It kind of is, but you can tell there's something a little bit off. And another thing is also when she's talking and when she's speaking to the camera, it also doesn't look natural, it looks pretty off. So, I'm pretty disappointed in Cling on this one. I thought it was going to be a lot better. So, I am going to go ahead and drop this one down to a four out of 10. All right, next up let's check out Omni Flash. >> I've been using this for about a week now, and honestly, I didn't expect to like it this much. >> Okay, so Omni Flash was actually really good. The first part was a little bit off the way that it cut through a few different scenes, but after she started talking, it was actually pretty straightforward, it was pretty natural, everything looks good. And I do like the extra little detail at the end where she goes to turn off the recording. So, overall, Omni Flash did a lot better than I expected it to. And I am going to give it a eight out of 10 because there's something off about it, which I can't really just pinpoint. I mean, besides the bottle because the text on the bottle, the logo, it is warping, doesn't look too good. All right, and now let's check out so far the king out of all of them, SeaDance 2.5. >> I've been using this for about a week now, and honestly, I didn't expect to like it this much. >> So, one thing that I didn't like was right off the bat is the text on the product itself. It's off. It doesn't really look too good. The music in the background looks nice. The way this is looking at the camera, it kind of reminds me of somebody that's looking at themselves when they're looking at their phone rather than looking at the actual camera. So, it does look pretty natural. I do like that part. >> I've been using this for about a week now, and honestly, I didn't expect to like it this much. >> The speech was pretty good. The lip-syncing was good. Overall, everything did look good. It was the bottle that kind of threw me off a bit, but I am going to give Seedance an eight out of 10 because it was really good. Now, let me actually give you a little bonus on this, and we're going to throw in Create UGC. So, I did talk to you about a little bit about this earlier, and Create UGC is simply a platform that's similar to Higgsfield what we're using now, but it's made specifically for e-commerce, and it's made for AI UGC videos. So, using this is actually very simple. It's a lot more straightforward than Higgsfield is in my opinion. And the way this works is you just sign up for an account. And once you sign in, you simply go ahead and click up here where it shows you create a new video. And you're going to have a few different ways to create that video. So, you can click on the product reaction, the product in hand, or an avatar and script. And the avatar and script is simply just a non-product video. You can have somebody just talking. And you also have a character swap, which is amazing because that means you can make multiple videos with the same elements and swapping out the speaker or the person that's talking and presenting. So, for this, what we're going to do is we're going to go ahead and do a product in hand. And as you see here, we do have all of the different models that you're going to be finding on Higgsfield as well. So, you have access to Kling, Veo, Seedance 2.0 and 2.5, and Kling 3.0 Omni. So, just as an example, let's create two different videos. We're going to be creating one with Kling and one with Seedance, the lowest and the highest scoring models from Higgsfield that we just used. Now, in order to generate these videos, it's actually really easy. All you have to do is go ahead and choose your model, click on continue, and then you're going to paste in your product link. So, what I'm going to do is I'm going to take the link from Amazon. I'm going to put it here where it says paste your product link and continue. Then it's going to give us our product description, the name, and it's even going to auto-fill all of the different assets for us. So, it already has different images that we can choose from. So, we don't have to generate anything on the side. We can choose our different video settings. I'm going to leave these as the default for now. And now I'm going to choose my creator. I love this part because here you can actually choose who's presenting your product. So, this person actually looks perfect for this type of video. Let's go ahead and click on next. I'm going to leave the default voice. We can go ahead and change that if you want to, but again, I'm going to leave mine as default. And then it generates our scripts. So, it's going to generate a funny, a selling, and a professional script. For this, let's go ahead and stick to the professional. Click on next, and now we just generate the video. So, we're going to do the same thing again for product reaction this time, and we're going to use C-Dance 2.5 for this. All right, so this is what we received for the product in hand. >> Experience the clinical efficacy of Neutrogena Collagen Bank. Featuring patented micro peptides and 15% vitamin C, this serum visibly improves skin tone and elasticity. Invest in your skin health and order your supply online today. >> So, this was actually very good. It was very straightforward. She was simply talking to the camera naturally. The voice didn't feel AI. It probably had a little bit of strange echo to it, but that's something that can easily be fixed with a little bit of editing. If you even decide to fix that. The way that the bottle is presented, the text on the bottle, everything was very spot-on, and nothing was being warped or even sounded awkward. So, this was really good. The only thing that I wasn't 100% on was a little bit of the sound, but besides that, I think that this came out absolutely amazing. Just add a little bit of music in the background, something soft, something calm, and you have perfect 10 out of 10 on this. So, clicking on create UGC for AI UGC ads did so much better than it did on Higgs Field." So, with this it's actually going to be getting a 10 out of 10. Now, let's check out how Sea Dance came up. For this one, she's not going to be holding the product in hand. She's actually doing a reaction to it. >> Experience the clinical efficacy of Neutrogena Collagen Bank Serum. Formulated with patented micropeptides and 15% vitamin C, it visibly firms and brightens your complexion. Invest in your skin and visit our website today. >> All right, this was hands-down again flawless. Like I think everything came out great. The only thing that I did not like was at the end how the serum is actually just hanging on top of her shoulder. But besides that, everything was great. And I love the fact that it added the serum actually coming out of the dropper. That was a very nice touch. So again, for this I'm actually going to give it a nine out of 10 just because of the fact that she has the bottle on her shoulder. But if it wasn't for that, perfect 10 out of 10. It even had a little bit of background music. Now, if you do want to check out Create UGC for yourself, it's actually a lot cheaper than all of the other models. I'll show you in just a second how much everything comes out to. You can get started at createugc.ai. The link to it is going to be down in the description below. All right, next up we have cinematic. The prompt that we're using is "The detective looks down at the old photograph and then notices someone watching him from across the street. He slowly looks up as a car passes between them, makes it feel tense and like a scene from a serious crime movie." So for this, let's start this time with Omni Flash. So we'll just switch it up. >> I know you are still [music] out there. I can feel you watching. >> Okay, so the first thing that caught my attention is the fact that it's raining and it looks like it's raining on the sky, but then when you look at the floor, it doesn't look like there's any raindrops that are going down. Besides that, he's looking at the back of the photograph instead of the front. I mean, he could be reading something, maybe, maybe not. The lighting overall is very cinematic. It's very dark, but I feel like it's missing some contrast. It's missing that little extra oomph. And his eyes had a weird little red glow to them. Wasn't too sure what happened there. >> Feel you watching. >> The voice overall just doesn't sound natural. It sounds pretty robotic. It sounds kind of AI-ish. And then after the car passes by, the movement of I guess what you would call the bad guy is not natural. The way that he moved backward looked a little bit awkward. And then just to top it off, the text is [snorts] not legible. So, everything around here and you can't really read anything. It says raft shigger or something, and then the rest of the letters just aren't actual letters. So, I am pretty disappointed in Omni Flash on this one. So, as far as cinematics for Omni Flash, I am going to be giving it a three out of 10. And that's actually me being nice. All right, so let's check out Kling 3 now. All right, so Kling actually had an amazing output, and I really liked how it started off. Overall, the vibe is very dramatic. It's very cinematic. It has a lot of contrast unlike on Omni Flash. It has some very nice darks. The reflection from the diner on the floor looks good. The sound effects on the zooming, everything is very smooth and it's working really well. And I really did like the reveal at the end when the guy shows up in the back. He shows up right after the car passes by. That transition is very nice, and it was very smooth. Kling for this, I think, just absolutely knocked it out of the park. So, as far as the vibes and the feel, very cinematic. Again, everything was on point. So, I'm going to give this one a 10 out of 10. I thought it was great. All right, let's check out Flux next. >> [music] >> Okay, so Flux worked pretty well at the beginning. I did like the headlights shining on the bottom of his coat. That pretty much just shows the car starting to drive towards him. The music was a little bit weird, but then when he turned around, that's kind of when it started to go downhill. The car is driving on a weird way. It just felt strange the way that it was going, as if it was going to jump onto the sidewalk. And the reveal of the guy in the background was a little bit off. And then just the fact that the detective turns around after he sees him and after the car passes by as if nothing happened, that throws it off, it started off good. It kind of derailed after that. So, maybe with a little bit of editing and a few extra shots that are generated, maybe we could fix it. So, I'm going to give it a 6 out of 10 for this. All right, and last in this batch is going to be C-Dance. So, everything was pretty good up until the car moved behind him. At that point, it kind of just threw everything off. So, C-Dance did pretty good in terms of the cinematic quality of it, but as far as making sense, it wasn't 100% spot on. So, I did like it, but I wasn't 100% on it. So, for this, C-Dance 2.5 is going to get a 7 out of 10 for cinematic. Looks like the king was dethroned by Kling. That's the thing, Kling was actually made for cinematic quality and cinematic shots. So, it makes sense that it won this round. All right, next we have lip sync and speech. So, one thing that I've noticed a lot of AIs issues with sometimes is making speech sound realistic or even look realistic. Now, for the lip syncing, we're going to be giving it a pretty long script. And for this, we're going to be using is, "She looks at the phone and says, "Wait, are you serious? I thought you said we were meeting tomorrow, not tonight. Now I have to change all of my plans and find someone to take care of the baby." She sounds surprised at first and then slightly annoyed. Match your mouth movements closely to what she's saying. Someone else is responding to her from the other side of the phone line and we're starting off with Flux. >> I'm sorry. I thought I told you it was tonight. >> Wait, are you serious? I thought you said we were meeting tomorrow, not tonight. Now I have to change all of my plans and find someone to take care of the baby. >> I know. I'm sorry. Can you still make it? >> Okay, so Flux actually did really good on this because the entire time that she was speaking and as she was moving around, as she was picking up the pace as she was speaking, it was all very smooth and the voice kept up with the lipsync. This I think was spot on. So Flux I think is going to get its first perfect score of 10 out of 10. I thought this was great. All right, next let's pull up Cling. >> Wait, are you serious? I thought you said we were meeting tomorrow, not tonight. Now I have to change all of my plans and find someone to take care of the baby. >> Okay, so for this one it kind of worked. At the beginning it actually did not match up at all. Then it kind of did catch up and then it lost it again. The beginning was a very awkward pause and nobody was even talking on the other side of the phone. So unfortunately Cling for this one did not really do too good. I'm going to have to give Cling for this a three or four out of 10. I'm going to stick to a three out of 10 for this one. All right, now let's check out OmniFlash. >> Wait, are you serious? >> Yeah, it's tonight. >> I thought you said we were meeting tomorrow, not tonight. No, check the text. Now I have to change all of my plans and find someone to take care of the baby. >> I'm really sorry about that. >> Okay, now for this one the lipsync was pretty good. It did look kind of AI-ish throughout the entire video, just the way that they were talking and the way that the video was structured. But another thing that I noticed is that the person on the other side of the phone started off as a man and then later on went to a female voice. So that was not consistent. Now as far as the lip sync on the woman that's talking, again, it was good, but there was something a little bit off about it. It did have that AI feel to it. So, I'm not going to give it a perfect 10 out of 10. I want to give it a between a six to seven out of 10. So, let's stick to a six out of 10 for this one. All right, and the last one for the lip sync is going to be SeaDance 2.5. >> Wait, are you serious? I thought you said we were meeting tomorrow, not tonight. Now, I have to change all of my plans and find someone to take care of the baby. >> I know. I know. I'm really sorry. It just came up. Can you make it work? >> All right, so the way that she talks is actually pretty good. She did have kind of a strange expression in there, and part of the video kind of makes it feel a little bit AI. Nothing too crazy. The background does make it look AI though, because the text on the stores don't match, or it's not real text. But, that's not the point of this. That's just me being picky. Since we're scoring strictly on the lip sync, it was good, but there was something a little bit off about it. It still felt AI. So, I'm going to give it an eight out of 10. All right, the next one is going to be prompt adherence. And that's simply going to be how well the model can follow our prompts without a reference image. So, for this we have a woman wearing a yellow raincoat walks into a small laundromat carrying a red umbrella. She closes the umbrella, places it inside the blue bucket by the door, and walks over to the third washing machine. That's pretty detailed. A black cat is sleeping on top of the machine. She drops one green sock, notices it, picks it up, and places it inside the washer. So, for this let's go ahead and start with SeaDance. Okay, so everything was pretty good. There was a cat at the beginning. It wasn't necessarily on the third washer. And then when she walked, she walked to the third washer from the wall in front of her. I thought she was going to walk to the third washer that was on her left. But, overall it was pretty good. It was convincing. So, I'm going to give it an eight out of 10. It wasn't perfect, but it was good. Now, let's check out Kling. Okay, so Kling did not do a good job. The way that the umbrella closed was inside out. It was really awkward. It obviously was wrong. And she didn't go to the third one. She actually went to the fourth machine. And when she picked up the sock, it kind of just raised on its own. She used some telekinesis for that. It did not work. So, as far as a prompt adherence, it did follow what we told it to do, but it didn't do it [clears throat] in the right way. So, because of that, we are going to drop it down to a four out of 10. All right, next up, let's check out Flux. Okay, so it didn't start off very well. She looked like she was exiting, and then she just closed the door. The scene completely changed, and then she put the umbrella in the bucket. But then, she didn't go to the third washing machine. She actually just chose the one that was in the middle. And if I count, it's actually the eighth washing machine. The one thing that I have noticed is consistent is going to be the cat and the green sock throughout all of the videos. So, Flux wasn't really my favorite. I'm going to give it a five out of 10. And then finally, we have Omni Flash. So, the moment that she walks in, when she puts down the umbrella, when it closes, it's off. It doesn't look real, and it was pretty awkward. Another thing that was weird on that part was the way that the clothes transferred from one arm to the other. Now, in this one, she actually dropped the sock rather than finding it on the floor. And again, we're not on the third washing machine. So, Omni Flash is going to have a pretty low score, kind of like the others. We're going to be giving it a four out of 10. All right, now the last test that we're going to do is going to be motion quality. Now, for motion quality, this is very important because this is something that is always a problem. When something moves too fast in an AI video, it typically tricks the model, it makes it confused, maybe, but it doesn't know where something goes. So, it starts to guess and it starts to hallucinate. So, for this, the prompt that we used was the person jumps rope quickly, alternating their feet, and moving from side to side. They cross their arms for one jump, uncross them, and continue with fast, tiny jumps. The rope keeps spinning without stopping, and the camera slowly moves around them. And the important part here is keeping the movement natural. So, let's start with Omni Flash. Okay, so Omni Flash actually did really good. The way that the jump rope was moving, the way that he was moving, everything looked very natural and very clean. And even in the background, you can see somebody recording him. That's absolutely crazy. That's some pretty high detail on that. So, Omni Flash for this one, I think absolutely killed it, and it's going to be getting a 10 out of 10. Now, let's pull up Kling. All right, so Kling started off a little bit okay. It looked like the only thing that was spinning on the jump rope though was just the two sides. It maybe could have been just motion blur, like it was going too fast, so you couldn't really see it too good. But as he kept going, as he kept moving, his face kind of started to warp a little bit, it didn't look too natural. And then it really lost it when he crossed his arms. The rope got stuck behind him, he kept going, and it just kept going as if he was actually jumping rope. So, Kling for this one did not do too good. I am going to be giving a four out of 10 because at first it did kind of work out, but it just lost everything afterwards. All right, so next up, let's check out Flux. >> Okay, so Flux was doing pretty good again up until point where he crossed his arms. When he crossed his arms, it kind of lost it. He wasn't really doing the jump ropes anymore. It looks like the jump rope was going from side to side rather than behind him. And with the motion, his face kind of started to blur. So, everything was good up until that part. It did do better than Kling though, so I am going to be giving Flux a six out of 10. All right, and last but not least, Sea Dance 2.5. Now, Sea Dance did really, really well. I can't really keep up when he crossed his arms. It did look really good. I thought maybe something looked awkward for a bit, but I don't think so. Everything actually came out really clean. And then at the end when he really started to pick up the pace, everything was keeping up really well and it was looking natural and real. There was no warping. Everything was pretty spot on on this. So, honestly, when it comes to Sea Dance, I'm going to give it a pretty high score. It's going to be getting a 10 out of 10. All right, so quick recap. In terms of realism, for this, our winner was hands down Sea Dance. The way that the barista gave the coffee and everything was working was hands down flawless. Now, unfortunately, on the flip side, Omni Flash and Kling were actually not that good with some strange effects happening. Then as far as our character consistency, Sea Dance again took the gold. Sea Dance is one of the most powerful models to date, so it's no surprise that it is going to be scoring higher than some of these other models on some aspects. The way that the woman was holding the record, looking through everything, everything was perfect. Everything stayed consistent throughout the entire video. And on the flip side, the lowest scoring to this was Flux. Everything that happened in that video was to me really off. It didn't really work. So, for character consistency, hands down was the winner. Now, for UGC, OmniFlash did surprisingly well. And C-Dance, of course, also did very well, too. Now, Kling, unfortunately, did not do very well at all in this case. But when we introduced Create UGC and Create UGC's programming of Kling, it came out flawless. It was absolutely perfect with a 10 out of 10. And same thing for C-Dance. C-Dance was also already pretty strong as it was when you include all of the aspects from Create UGC, you know, just simply the fact that Create UGC was made to be able to use these models for AI UGC, it upgrades all of these models instantly for this type of content. And again, C-Dance for this came out at a nine out of 10, as well. So, hands-down, the winner for this, I do have to give it to Create UGC because both generations, even the worst model using Higgs Field, was really good. Now, as far as our cinematics, OmniFlash was the worst one. The contrast wasn't there. It didn't look dynamic. It didn't look dramatic at all. And on the flip side, Kling was perfect. But Kling is made for cinema. It's made for very dramatic scenes. So, it makes sense that it won. Now, the lip sync. This was actually one of my favorite ones because this is something that's very important. If somebody sees a video and the voice isn't synced to the lips, the speaking is off, they're automatically going to be turned away. I'm one of those people. I see the voice doesn't match the mouth. I can't stand it. So, for this, unfortunately, Kling was actually the worst one with a three out of 10. At the beginning, it wasn't very good. It got a little bit better, but then it went back down, and it just wasn't good at all. It's not something that would be usable. On the flip side, Flux was hands-down perfect with a 10 out of 10. Now, prompt adherence. So, no reference image at all. For this, they were all kind of off. C-Dance, though, was hands-down the winner, but it still was not perfect. And we did end up giving it eight out of 10. And then, for the motion. This is also, in my opinion, just as important as the lip syncing. And OmniFlash completely took it, blew everything else out of water with a 10 out of 10. Everything on it was perfect and super smooth. Nothing blurred out, nothing got awkward, nothing got lost in that fast movement. And same goes for SeaDance. SeaDance also got the perfect 10 out of 10. But on the flip side, Kling didn't do very well. Everything started to get lost, it got jumbled, the movements were awkward, and it just didn't make sense. Now, as a general average for the different models that we used, Flux 3 used about 66 credits, Kling used about 72, Gemini Omniflash used about 30, which was actually the cheapest, and SeaDance 2.5 came in at the most expensive at 78 credits. Now, the plus pricing for HeyGen does come around $49, and after we do the math, Flux comes out to about $3.23 per generation, Kling is about $3.53, Gemini is $1.47, and SeaDance is $3.82. So, overall, it's not a bad price, but as you start to use these more and more, as you start to generate more videos, especially with the better models, like let's say SeaDance 2.5, it is going to start to get very expensive. But that's actually where Create UGC comes in, because Create UGC actually averages out to just $1 per generation for all of the different models that we have. Whether you're using Kling or SeaDance 2.5, you can get your video generations for as low as $1 per video. So, if you're specifically looking to create AI UGC videos for your brand or for, let's say, an e-commerce store that you're running, Create UGC is going to be your go-to. And remember, the link to that is going to be down in the description below. So, I want to know which one of these was your favorite model, which one of these do you think is best [music] for what? Let me know down in the comments below, I'd love to hear your thoughts on this. And if you do want to learn how you can connect HeyGen field and Claude to create AI UGC videos like this, but directly through Claude, then check out this video right here.