WEBVTT 00:00:00.000 --> 00:00:02.260 shooting out. I believe that was something I've seen. 00:00:05.200 --> 00:00:11.240 The trees are quite reminiscent of a poo. Deep Seek V4 Pro has just been released in 00:00:11.240 --> 00:00:15.620 general available version, which also happened with the V4 Flash model back 00:00:15.620 --> 00:00:22.260 about 12 or 13 days ago on 0731. Now that model, when tested, proved to be a significant, 00:00:22.460 --> 00:00:26.900 significant increase in capability over the preview version that was previously around. 00:00:26.900 --> 00:00:31.620 and today we're going to be finding out if the same can be said for the newly updated Checkpoint 00:00:31.620 --> 00:00:37.900 general availability for V4 Pro dubbed 0813. So before we get into it do feel free to subscribe 00:00:37.900 --> 00:00:43.240 so we can hit 70k and then 100k with the plaque and with that basically there's not a lot to say 00:00:43.240 --> 00:00:47.500 about this right now because it seems to have been more or less silently released. There's no 00:00:47.500 --> 00:00:52.000 official announcement post yet on X which I do believe is generally where we would see this 00:00:52.000 --> 00:00:56.140 but they do mention here just in the website a couple references to it where 00:00:56.140 --> 00:01:03.500 If you use this model identifier right here for V4 Pro, you are being routed to the new V4 Pro 0813 version. 00:01:03.820 --> 00:01:12.000 Additionally to that, it is also live on open router and this makes me feel comfortable that it is actually genuinely released and this is the model we're going to be using. 00:01:12.160 --> 00:01:19.240 So from the bit of information we have right here, aside from just the model identifier now being today's date, depending on what locale one is in, 00:01:19.240 --> 00:01:25.900 we can see that the model has a million context window with a maximum output length of 384k, which really is a lot. 00:01:26.140 --> 00:01:31.520 Now, something that can sometimes happen with DeepSeek, at least historically, is there may be a lot of thinking. 00:01:31.940 --> 00:01:38.140 Now, this does support selectable thinking modes, as you can see if we do click on this, so it's possible to kind of tone that down. 00:01:38.140 --> 00:01:45.360 But in the interest of seeing the maximum potential performance, all of our coding tasks for today are going to be run on this maximum effort. 00:01:45.820 --> 00:01:53.760 Now, I want to mention the following with a grain of salt because I cannot definitively confirm that this is ground truth, but I would imagine it probably is. 00:01:53.760 --> 00:01:58.740 There were some references and photos floating around which apparently came from a group related 00:01:58.740 --> 00:02:04.000 to DeepSeek. I believe they said it was a WeChat group about DeepSeek or something. And if this is 00:02:04.000 --> 00:02:08.160 to be believed, we can see that these would hypothetically be the benchmark results for 00:02:08.160 --> 00:02:14.600 DeepSeek V4 Pro 0813 or the general availability model. And basically my main interest here is 00:02:14.600 --> 00:02:20.080 just seeing the leaping capability over the V4 Pro preview to now because I was initially 00:02:20.080 --> 00:02:26.160 disappointed with the v4 pro preview as well as the v4 flash preview they felt very undercooked and 00:02:26.160 --> 00:02:31.200 probably because they were named preview that's actually like fair it makes sense but again i 00:02:31.200 --> 00:02:36.340 want to reiterate the leaping capability between flash and then the 0731 general availability 00:02:36.340 --> 00:02:42.300 version of flash was absolutely massive this is like the best local model right now for like a 00:02:42.300 --> 00:02:49.200 dual spark setup so i mean in a niche you know but still if that leaping capability remains then what 00:02:49.200 --> 00:02:55.660 we can see right here, just in this deep SWE score increase from 12.8 to 62.7 would then 00:02:55.660 --> 00:03:01.040 hypothetically put this above things like Opus 4.8 and make it very, very competitive. Being 00:03:01.040 --> 00:03:06.220 especially that this is an open weight model or will be, that's also extremely exciting because 00:03:06.220 --> 00:03:12.700 this is not like horrifically gigantic. I guess when considering things like Kimi K3 or Quen 3.8 00:03:12.700 --> 00:03:18.620 max, it's still big at 1.6 trillion parameters. It is a mixture of experts model with how many 00:03:18.620 --> 00:03:24.060 active. So if we do estimate that this is going to be the same size as the preview version of V4 Pro, 00:03:24.180 --> 00:03:30.080 it's a 1.6 trillion parameter mixture of experts model with 49 billion active. So it is still very, 00:03:30.200 --> 00:03:35.640 very large, but it's not as gigantic as like the closing in on 3 trillion parameter models. So 00:03:35.640 --> 00:03:40.800 additionally to that pricing, and this is something that we need to basically take just as it is right 00:03:40.800 --> 00:03:45.120 now, because they mentioned and have been for quite a bit of time that their API pricing is 00:03:45.120 --> 00:03:50.280 going to be significantly increased soon. Now significantly is relative to what it currently is, 00:03:50.340 --> 00:03:56.220 which is like very, very, very cheap. So right now pricing per million input tokens for DeepSeek 00:03:56.220 --> 00:04:04.800 V4 Pro is 4.35 cents. So, okay, now I know what this is, but like last time I messed up on this 00:04:04.800 --> 00:04:11.720 one and I said it was two cents and it wasn't, and then I got roasted. So this is 43 cents per 00:04:11.720 --> 00:04:17.060 million input tokens and 87 cents per million output tokens is the current price. Now, if this 00:04:17.060 --> 00:04:22.360 gets like a significant increase, say it goes up like five or 10 times, it would still hypothetically 00:04:22.360 --> 00:04:27.480 be competitive. So it's all relative to what it currently costs. And with that, I am going to be 00:04:27.480 --> 00:04:33.700 using this from within Codex. When we tested V4 Flash, the general availability model on the 31st 00:04:33.700 --> 00:04:38.640 of July, I did also use it through Codex because I believe they had suggested to do so. And the 00:04:38.640 --> 00:04:44.280 performance was very very good I was happy with it so I'm going to be using Codex as well for this 00:04:44.280 --> 00:04:52.060 now I should also mention that they did release in beta a desktop app for Codex or ChatGPT so I 00:04:52.060 --> 00:04:56.220 also have it routed and working through this desktop app as well and it is doing a skate game 00:04:56.220 --> 00:05:00.540 for us in C++ I would imagine that will take quite a bit of time so I did just want to start 00:05:00.540 --> 00:05:04.840 it and verify this app was actually working properly on this system and it does seem to so 00:05:04.840 --> 00:05:11.400 keep in mind any mention of GPT or codex here is routed directly to DeepSeq on max reasoning effort 00:05:11.400 --> 00:05:17.580 for the 0813 model. And with that, let's take a look at our, okay, so apparently the OpenAI 00:05:17.580 --> 00:05:24.860 beta application set itself to be the default for all file types. That's infuriating. All right. 00:05:25.160 --> 00:05:31.540 All right. Very Quen 3.8 aesthetic. I'm going to turn sound on because I did see some mentions of 00:05:31.540 --> 00:05:38.340 it just talking about like sounds. All right. It starts in our notepad. Can we resize this? 00:05:38.680 --> 00:05:45.560 All right. Acceptable. Minimize. Unminimize. Good. The big one. Is there a right click? Yeah, 00:05:45.620 --> 00:05:52.660 there is. There is also the correct date and time in my locale in the bottom right. And there is 00:05:52.660 --> 00:05:56.840 what I would imagine here to probably be the special feature. It is the holographic desktop 00:05:56.840 --> 00:06:02.180 feature so we will save that for last. Let's start out by looking at the files app which I don't 00:06:02.180 --> 00:06:06.480 think I've seen a files icon that looks like this before. All right we have some nice transparency 00:06:06.480 --> 00:06:11.680 to the application backgrounds and definitely a nice synthwave theme which I do personally like. 00:06:11.820 --> 00:06:15.520 So unfortunately it seems our fake file system is just static and we can't browse through it. 00:06:15.520 --> 00:06:20.240 Nonetheless can we create new things in it? Okay we can. Can we remove the new things we created? 00:06:20.240 --> 00:06:27.660 we also can can we remove things that were there by default yeah we can all right cool that makes 00:06:27.660 --> 00:06:33.600 me a little happier and then up would just go there oh okay not bad decent bit of immersion 00:06:33.600 --> 00:06:38.660 at least in terms of what directories would exist were this a real computer notepad we did kind of 00:06:38.660 --> 00:06:42.880 see because it opened with this just to give us some information the only thing i like to see is 00:06:42.880 --> 00:06:48.320 will it save it as a text file on the system no such directory didn't i delete the documents 00:06:48.320 --> 00:06:53.700 directory. Hold on a second. That's okay. Oh, it remembered that I did that. I might have messed 00:06:53.700 --> 00:07:00.820 that up. So that's going to be, that's on me. Okay. This right here is just not right. It should not 00:07:00.820 --> 00:07:07.060 have a like visible text box that happens. Sometimes help did give us things. Let's just 00:07:07.060 --> 00:07:17.640 do Neo fetch. That's actually pretty well done right there. Negula or nebula, nebula, 00:07:17.640 --> 00:07:21.700 still, that's actually alright. I'm just being kind of mean. 00:07:22.220 --> 00:07:23.880 Memory, this is fine. 00:07:24.880 --> 00:07:27.640 Oh, it actually put a bit of cleverness in there. 00:07:28.240 --> 00:07:30.760 1.44 megs out of infinity, this is fine. 00:07:31.080 --> 00:07:33.300 Isn't that how much an old school floppy disk would hold? 00:07:33.840 --> 00:07:35.560 I don't know, because those are older than me. 00:07:36.020 --> 00:07:37.580 Next up, calculator. 00:07:37.580 --> 00:07:41.740 Okay, so 47 times 8. 00:07:44.760 --> 00:07:45.200 376. 00:07:45.620 --> 00:07:47.040 I haven't done one of these in a while. 00:07:47.040 --> 00:07:49.980 divided by 6. Alright. Settings. 00:07:52.400 --> 00:07:55.120 Definitely seeing a bit of like Quen 3.8 00:07:55.120 --> 00:07:58.680 in here just in terms of some of the style here. What is that? 00:07:58.800 --> 00:07:59.780 And why do I like it? 00:08:01.505 --> 00:08:10.685 aurora oh okay sunset city this no one would understand that reference this looks like 00:08:10.685 --> 00:08:15.125 something one of my friends in high school was making as a game project it actually looks like 00:08:15.125 --> 00:08:23.125 a ton like the splash screen for his project we have customizable accents as well difficulty easy 00:08:23.125 --> 00:08:29.205 normal and hard all for games okay i'll probably set it to easy as you may be familiar with my 00:08:29.205 --> 00:08:37.045 quick to anger game playing style next up let's try our games the speaker is on so city heist 00:08:40.885 --> 00:08:42.405 oh you absolute 00:08:45.125 --> 00:08:52.305 Okay, so you have to press enter to start and not click the button. 00:08:52.305 --> 00:08:53.305 Please shift gears. 00:08:53.305 --> 00:08:56.305 Oh, okay. 00:08:56.305 --> 00:08:57.305 Okay. 00:08:57.305 --> 00:09:02.445 All right, it's... 00:09:02.445 --> 00:09:07.445 Oh, cool. 00:09:07.445 --> 00:09:14.125 Oh, okay. 00:09:14.125 --> 00:09:24.805 troubled but the sounds for like this smack are kind of funny like all right 00:09:24.805 --> 00:09:29.165 and we wrecked and we only grabbed a hundred out of a thousand but it did 00:09:29.165 --> 00:09:32.845 work so keep in mind we need to press enter to play these games and not hit 00:09:32.845 --> 00:09:48.105 the button that says launch. Oh, all right. Do I go through these? Yeah. I just assume based on 00:09:48.105 --> 00:09:52.505 the sounds. I wonder if that's how like dogs are since they don't understand language, but like 00:09:52.505 --> 00:09:56.465 that to me seemed like a positive reinforcement sound. So it makes me want to do it more. 00:09:56.465 --> 00:09:59.905 And I think we did just discover something about how dog screens work. 00:09:59.905 --> 00:10:08.285 okay the score I don't know that we need that many decimal points just to show a score I think 00:10:08.285 --> 00:10:15.025 maybe just rounding up to 734 or down would have been acceptable but I applaud it for its 00:10:15.025 --> 00:10:20.825 intricacy and then oh read me okay so our special feature is something like holographic 00:10:20.825 --> 00:10:25.665 and I would imagine having a couple apps open just when we trigger that is probably the move so 00:10:25.665 --> 00:10:31.085 oh yeah all right so it just puts like an interesting card effect this is what it looks 00:10:31.085 --> 00:10:36.245 like this looks like an apple vision pro like app demo this is the way they look from inside the 00:10:36.245 --> 00:10:44.805 vision pro headset just if you have used one you may resonate with what i'm saying okay overall 00:10:44.805 --> 00:10:50.125 it was fine everything worked nothing really blew me away but again that may be a limitation of the 00:10:50.125 --> 00:10:56.165 test and less of a limitation of the model. So alas, it's kind of neat. We will continue. 00:10:56.405 --> 00:11:02.125 So our skateboard game is still going. And again, I'm using this with the newly released beta version 00:11:02.125 --> 00:11:07.005 of the chat GPT desktop app. I would imagine this will take quite a bit of time. So let's check to 00:11:07.005 --> 00:11:11.725 see just like, oh no, because we already have looked at the cost and keep in mind, this is 00:11:11.725 --> 00:11:18.585 partially for the skate game as well as the browser OS. So 20 cents total across two separate tasks, 00:11:18.585 --> 00:11:25.685 one which is still ongoing all right so after 68 minutes and 24 seconds skate nyc is complete 00:11:25.685 --> 00:11:31.145 so oh you idiot you spoiled it with these screenshots and you don't even have vision 00:11:31.145 --> 00:11:38.925 i'm infuriated nonetheless now that we know what the game looks like i had to stop myself 00:11:38.925 --> 00:11:44.785 from saying something a little more colorful let's just take a peek at the game all right so 00:11:44.785 --> 00:11:55.885 here's our NYC skate game should be in outputs right yes all right it's not quite as good as 00:11:55.885 --> 00:12:01.865 the quen 3.8 model but that model does have 800 billion more parameters 00:12:01.865 --> 00:12:09.385 does that say NYC skate on the ground it's inverted but it does and it says established 00:12:09.385 --> 00:12:18.865 1999 all right sounded like the beginning of the Van Halen song I don't know what exactly is all 00:12:18.865 --> 00:12:24.745 right let's just maybe not do that same thing there are sound effects though I don't normally 00:12:24.745 --> 00:12:32.305 hear sound effects can we actually move beyond all right not quite I will say it's not wonderful 00:12:32.305 --> 00:12:39.325 that's actually kind of cool though the way it like stylized the game title nice nice 00:12:39.325 --> 00:12:46.625 well we don't need to do the drum kit sim the trees are 00:12:46.625 --> 00:12:55.765 quite reminiscent of a poo if i were to be honest but all right the legs are like kicking 00:12:55.765 --> 00:13:06.545 this may be the funniest skate game result we've ever seen which is you know something all right 00:13:06.545 --> 00:13:13.045 we have a street sign there the text is a little sideways as are the building faces this is 00:13:13.045 --> 00:13:18.185 definitely a troubled result however i do like the water fountain shooting out i believe that 00:13:18.185 --> 00:13:27.165 was something I've seen. I've seen that in, I think the Quinn result. Now here's what I would 00:13:27.165 --> 00:13:32.365 like to just quickly do for some reference. This is the result we received for the C++ skateboard 00:13:32.365 --> 00:13:38.825 test from the preview version of DeepSeek V4. We can see right here, I'm going to turn the speaker 00:13:38.825 --> 00:13:45.505 off. This is pretty bad. It was like so bad that I actually remember to go to this specific thing 00:13:45.505 --> 00:13:51.885 just to showcase it. So the leap in capability based off of what we see right here to what we 00:13:51.885 --> 00:13:56.865 just saw in the skate game that it had created for us, which we can just look at in these screenshots 00:13:56.865 --> 00:14:02.405 here, is still pretty significant. Next up, I'm giving it the subway FPS test. I've also added 00:14:02.405 --> 00:14:06.625 on to the end of the prompt. Don't worry about checking over the script. I will manually check 00:14:06.625 --> 00:14:10.645 it and if there are issues, I will report them back to you just because this will really like 00:14:10.645 --> 00:14:17.225 extend every single result by like two times in terms of like time-wise length just in the interest 00:14:17.225 --> 00:14:21.345 of something I could do visually in like five seconds and then be like hey this doesn't work 00:14:21.345 --> 00:14:26.605 so we're gonna do that just to maybe like keep things a little tighter all right so it did finish 00:14:26.605 --> 00:14:31.385 the subway fps game the problem is it just keeps trying to do like troubleshooting which 00:14:31.385 --> 00:14:36.705 it's not good at doing on this system for some reason all right let's check out our subway fps 00:14:36.705 --> 00:14:44.185 Montrand Avenue. All right. I'm going to turn the speaker on. So it gave us two options. We can 00:14:44.185 --> 00:14:53.185 either explore the station. Okay. It's, Hey, you know what? I think that is the best subway car 00:14:53.185 --> 00:14:58.485 that I've ever seen drawn in this test. And you're going to be like, well, it doesn't look that great. 00:14:58.485 --> 00:15:03.365 I know. It's just that they never really include these. And when they do, they're not really very 00:15:03.365 --> 00:15:10.125 good. So this actually, what was that? If that thing starts moving, I would have probably become 00:15:10.125 --> 00:15:16.645 more excited than would be reasonable for, you know, responding to such thing. This is very clean. 00:15:16.765 --> 00:15:20.925 It's simple. It's nothing that's blowing me away, to be completely honest with you. Queens Boulevard 00:15:20.925 --> 00:15:25.905 line. Okay, we have stairs. We have some lighting elements and things of the sort. We have ace, 00:15:25.905 --> 00:15:33.665 if you want to go that side. And then tunnels, like indicators for the train's trash bin, 00:15:34.425 --> 00:15:39.385 floating particles, more simple, and we can't go up the stairs, which is fine. We don't want to. 00:15:39.705 --> 00:15:46.245 Now, of course, the other feature that was included here is the FPS. So let's go back to 00:15:46.245 --> 00:15:51.165 menu and then enter the survival FPS. Let's see if it changes the lighting or anything. Oh my, 00:15:51.165 --> 00:16:02.925 that's a good weapon model does it how oh it does I think yeah it does it's like smooth now I don't 00:16:02.925 --> 00:16:10.905 see any enemies look at when we run it's like a waddle all right so uh the big thing I'm 00:16:10.905 --> 00:16:13.705 noticing here is there are no zombie enemies so 00:16:15.887 --> 00:16:18.507 Good reload. The waddle is just kind of funny. 00:16:22.827 --> 00:16:27.247 Weird. Hey, look, there are shell casings that actually drop. That's kind of cool. 00:16:27.247 --> 00:16:32.427 All right, this is like, yeah, it's kind of mid. So I'm going to give it a feedback piece. 00:16:33.187 --> 00:16:36.727 Now I stopped it because it was just like, it had finished and it was like, 00:16:36.907 --> 00:16:40.487 I can't use this in the headless environment to check. And at that point, it's easier to 00:16:40.487 --> 00:16:44.227 just check it and tell it what's wrong. So I'll give it that honest feedback. 00:16:44.387 --> 00:16:48.907 So supposedly our fixed and polished version of the Subway FPS is completed. 00:16:49.347 --> 00:16:55.347 Let's just take a peek at it. The sound is on. All right. Interesting. I think it swapped the 00:16:55.347 --> 00:17:02.487 positioning of the buttons. So, ooh, all right. Seven hostiles. It's showing me in the top right. 00:17:02.847 --> 00:17:10.407 I unfortunately just can't locate them. I would have expected it to have done better. And again, 00:17:10.407 --> 00:17:15.527 this is one where we could go back and reference the previous model of this. So basically the 00:17:15.527 --> 00:17:23.347 preview version, it was significantly less detailed than this was. Though unfortunately, 00:17:23.347 --> 00:17:28.887 we don't have enemies and given the amount of time it took as well as the kind of multiple times I 00:17:28.887 --> 00:17:33.267 told that there were no enemies I'm regrettably going to call this one. Next up I'm giving this 00:17:33.267 --> 00:17:40.107 a difficult task to create a 3D model of an RB26 inline-six engine. It must be something that can 00:17:40.107 --> 00:17:46.227 be 3D printed with little to no support and fit inside of it a tiny DC motor called an N20 motor 00:17:46.227 --> 00:17:51.347 because this is something that I've told it should like work with my RC car project. Actually I don't 00:17:51.347 --> 00:17:55.767 think that's specifically in this prompt but that's the whole point of this where this is it 00:17:55.767 --> 00:18:00.767 should understand and it will scale it properly to that motor so far I've run this test with a few 00:18:00.767 --> 00:18:04.707 intelligent models and I've never really received anything that was actually worth printing so 00:18:04.707 --> 00:18:10.287 hopefully sooner than later that will change so our 3d model task is pretty interesting because 00:18:10.287 --> 00:18:15.207 it's just doing a lot of research on like the actual proper orientation of specific items on 00:18:15.207 --> 00:18:20.307 this engine like where's the downpipe which is an exhaust component what side is it what is the 00:18:20.307 --> 00:18:25.927 intake side? What is the exhaust side? Intake, plenum, left or right? So I do like to see the 00:18:25.927 --> 00:18:30.307 amount of like intricate research that's going into the culmination of this model creation. 00:18:30.647 --> 00:18:36.427 And it would be really awesome for like us to get a model for this task that's actually printable. 00:18:36.507 --> 00:18:41.987 It would be the first one. So in one hour and 10 minutes, we received our 3D model response. 00:18:42.167 --> 00:18:47.647 This is a very, very intricate, just bit of text about the thing it made. Now, of course, 00:18:47.647 --> 00:18:52.567 the other part of that is actually just taking a look at what it looks like though once I remember 00:18:52.567 --> 00:18:57.267 what I named this folder we will be able to take a peek assembled display stl I would imagine that 00:18:57.267 --> 00:19:06.047 to be the case please be good I've really like I need a good one for these actually it's sideways 00:19:06.047 --> 00:19:16.167 son of a gun this did a really good job this is actually like so far the best one it's not 00:19:16.167 --> 00:19:23.267 perfect we have some sideways turbocharger issues though i mean this is all right let's let's peek 00:19:23.267 --> 00:19:27.367 at this in open scad there are a bunch of individual pieces for this so we also want to 00:19:27.367 --> 00:19:31.307 see if the engine that we told it needs to fit inside the little remote control car engine 00:19:31.307 --> 00:19:38.147 let's see if it has any hope of properly fitting research it has a research folder right here just 00:19:38.147 --> 00:19:45.587 showing like a bunch of images and stats for this motor very cool though and here we have some better 00:19:45.587 --> 00:19:55.667 preview renders. This is definitely well done. Turbo side, okay, underside, all right. It's not 00:19:55.667 --> 00:20:00.367 something that's going to really easily be 3D printable, although I guess I can't make that 00:20:00.367 --> 00:20:05.347 determination because this comes in multiple pieces. So it's not designed to be printed as 00:20:05.347 --> 00:20:10.587 we first saw it just in one lump like this. It's designed to be printed separately. I would like 00:20:10.587 --> 00:20:16.727 to see. Let's look now at just the single first piece, which is ideally where it would be placing 00:20:16.727 --> 00:20:24.227 the 3D printed motor that would go in it. Unfortunately, printing this without supports 00:20:24.227 --> 00:20:29.947 is just not an option in any potential orientation that we see right here, which is a bit frustrating. 00:20:30.107 --> 00:20:34.187 It almost seems like it has these posts and holes in it to be able to snap fin on other parts, 00:20:34.307 --> 00:20:39.807 which is cool. Okay. Yep. And I think those are definitely like attachment posts, which is pretty 00:20:39.807 --> 00:20:44.447 cool. This one would be able to be printed more or less without any supports if we were to 00:20:44.447 --> 00:20:51.167 put it flat. Then this should have the Nissan logo or at least a text Nissan written on it. 00:20:53.107 --> 00:20:59.347 Yep and this one's quite well done. No issues here. The oil pan. Oh no. Okay this is where we 00:20:59.347 --> 00:21:03.967 this is where we decide we're not going to be trying to print this. I'm not disappointed 00:21:03.967 --> 00:21:10.927 but I'm not super impressed either so I suppose I am lukewarm when it comes to judging this 00:21:10.927 --> 00:21:16.607 definitely if we just look at this as a render even forget about looking at the individual parts 00:21:16.607 --> 00:21:22.127 there because it's not really going to be feasible to print this it did a really decent job just 00:21:22.127 --> 00:21:26.827 assembling it the research that went into this I think is almost probably like I hate to sound 00:21:26.827 --> 00:21:31.527 like an AI here but like the story isn't the model it's the research that went into it that's 00:21:31.527 --> 00:21:38.427 generic like AI slob. It really did a lot of decent research here. I mean, did it find this 00:21:38.427 --> 00:21:44.887 and download it just to see? No. Okay, so this is something it maybe made or maybe probably found 00:21:44.887 --> 00:21:49.907 online just to be able to better assess what the size of this asset is. But it took a lot of care 00:21:49.907 --> 00:21:54.947 to ensure that what it tried to generate would have hypothetically fit the specific specifications 00:21:54.947 --> 00:22:01.347 that were denoted in the original prompt. So not half bad. I'm also going to give this a simple 00:22:01.347 --> 00:22:05.507 front-end design test where it needs to create the watch website. So it has to have a pretty 00:22:05.507 --> 00:22:10.987 cinematic panning effect over a 3D watch asset that it creates. I didn't specifically, okay, 00:22:11.067 --> 00:22:15.787 no, I did specifically denote that the watch needs to be a 3D model. So we'll see how competent it is 00:22:15.787 --> 00:22:21.627 at putting the entirety of this together. So in one hour and 30 minutes, we got our watch front-end 00:22:21.627 --> 00:22:27.347 result. It again seems to have a pretty substantial like just list of what was built here. So let's 00:22:27.347 --> 00:22:35.187 take a look and see how it is. Beautiful watch front end website. That's actually so close to 00:22:35.187 --> 00:22:42.527 excellent. The big problem is this, I would assume is the watch strap sticking out. This is 3D. So 00:22:42.527 --> 00:22:48.427 don't be fooled by its 2D-ishness look. It is actually 3D. It's slowly panning too. I really, 00:22:48.747 --> 00:22:54.567 okay. So a couple of things I do like the wooden background and then this like leather card or pad 00:22:54.567 --> 00:22:59.067 that this watch is placed on. Additionally to that, the face looks really good. The materials 00:22:59.067 --> 00:23:05.307 look very good. It's just like an overall, it's a high-end aesthetic, which is exactly what we go 00:23:05.307 --> 00:23:11.427 for when performing this test. Now, that's actually, that right there looks good. The, 00:23:13.267 --> 00:23:19.347 let's just sit and appreciate for a minute. Okay, yeah, that, maybe I could have done without that, 00:23:19.347 --> 00:23:25.147 but it's definitely quite something. And then look at the way that goes away as soon as we 00:23:25.147 --> 00:23:32.627 start to scroll down. Quiet luxury measured in seconds. Sunburst dial. Okay, it has like a 00:23:32.627 --> 00:23:37.807 material card here. The sapphire crystal is perhaps something I would have maybe omitted from this 00:23:37.807 --> 00:23:43.267 section, but the dial looks good. Two watches, one point of view. Please render them. Oh, look at that. 00:23:43.367 --> 00:23:47.487 That is sick. That really looks pretty good. I love the way they're placed on the table. 00:23:47.487 --> 00:23:52.527 really the big thing letting it down is just this leather slab is inhibiting our view of these 00:23:52.527 --> 00:23:58.527 beautiful straps this one is the stainless steel option this one is the gold option and priced 00:23:58.527 --> 00:24:05.387 accordingly i would imagine yeah although i would say the jump in price of like that is the 00:24:05.387 --> 00:24:09.787 negligible amount to get the gold one i'd probably go with this one personally oh cool and these ones 00:24:09.787 --> 00:24:11.907 we can freely rotate so that helps a bit 00:24:14.247 --> 00:24:18.907 watch hands, but that's okay. I'll forgive it for that. But these are pretty darn cool. 00:24:19.487 --> 00:24:26.947 These are well drawn, I would say, especially the numeral indicators. Look at the way that 00:24:26.947 --> 00:24:31.907 they actually have multi-material on them. So they're not just like one static material, 00:24:32.087 --> 00:24:37.087 it's gold and then black infill or silver and then black infill as well. So that's some attention 00:24:37.087 --> 00:24:42.887 to detail I do like seeing as well as the kind of brushed material look on the stainless steel and 00:24:42.887 --> 00:24:48.227 the gold. It's probably weird we're like assessing like a 3D watch, but it just shows like some 00:24:48.227 --> 00:24:52.287 capability. Okay, that looks pretty good right there. And then we have a clean footer as well. 00:24:52.287 --> 00:24:59.327 I like this. I'm happy with this. Even though it's like not perfect, it's darn close. So next up, 00:24:59.447 --> 00:25:03.547 just here from within Codex, and I can't show this full prompt because it has an API key. 00:25:03.767 --> 00:25:08.567 I'm giving this the full-on cinematic Steve the PC repair game test where it needs to create a 00:25:08.567 --> 00:25:14.047 cinematic game. It needs to create voice assets for these characters just by using an OpenAI API 00:25:14.047 --> 00:25:19.307 key that I've given it, and it needs to make a fun game that is enjoyable and playable. I will 00:25:19.307 --> 00:25:23.567 assess this is probably going to take north of an hour so we're just going to basically leave 00:25:23.567 --> 00:25:28.807 codex here in the background the beta desktop version of it to complete this task and we'll 00:25:28.807 --> 00:25:33.487 maybe work on some other things in the meantime so our steve the pc repairman cinematic game 00:25:33.487 --> 00:25:39.567 completed in basically 65 minutes so a little over an hour all right here's how to play i'll 00:25:39.567 --> 00:25:44.747 probably just do what it says here not assume that it's already running this in some server for us 00:25:44.747 --> 00:25:45.967 So, okay. 00:25:46.467 --> 00:25:48.527 Episode one, the Azure Palm job. 00:25:49.147 --> 00:25:49.707 Okay. 00:25:50.147 --> 00:25:50.947 Click to begin. 00:25:51.127 --> 00:25:51.987 Headphones recommended. 00:25:52.227 --> 00:25:52.947 We have a speaker. 00:25:54.647 --> 00:25:55.387 All right. 00:25:57.687 --> 00:25:59.387 My name is Steve. 00:25:59.447 --> 00:25:59.987 This is good. 00:26:00.167 --> 00:26:01.327 I fix computers. 00:26:01.587 --> 00:26:01.947 Well, okay. 00:26:01.947 --> 00:26:06.667 Most of my clients are old enough to remember when a hard drive was a luxury item. 00:26:07.387 --> 00:26:10.347 They pay what they can afford and I make it work. 00:26:11.007 --> 00:26:12.467 It is honest business. 00:26:14.007 --> 00:26:14.567 Oh. 00:26:14.747 --> 00:26:15.567 Thank you, dear. 00:26:16.067 --> 00:26:18.307 The screen kept telling me the date was 1980. 00:26:19.107 --> 00:26:22.447 And I thought, well, I have been time-traveling long enough. 00:26:22.447 --> 00:26:23.107 Fresh battery. 00:26:23.827 --> 00:26:24.907 Your computer now believes it is 2026. 00:26:24.907 --> 00:26:26.887 I will bring you oatmeal cookies on Tuesday. 00:26:27.007 --> 00:26:27.727 Same as the rest of us. 00:26:27.727 --> 00:26:29.867 And give Mr. Buttons a scratch behind the ears for me. 00:26:29.947 --> 00:26:32.567 Let me walk you out, ma'am. 00:26:33.907 --> 00:26:36.467 I'll be right with you, Mr. Thomas. 00:26:36.467 --> 00:26:38.687 The speech overlays itself. 00:26:38.707 --> 00:26:39.807 You are a good boy, Stephen. 00:26:40.447 --> 00:26:44.107 If that quiet man needs a computer fixed, you be patient with him. 00:26:44.747 --> 00:26:45.987 I like punctual. 00:26:45.987 --> 00:26:47.347 Door is unlocked. 00:26:47.347 --> 00:26:48.947 Coffee is warm. 00:26:48.947 --> 00:26:49.947 A man. 00:26:49.947 --> 00:26:50.947 What is the job? 00:26:50.947 --> 00:26:51.947 Victor Kane. 00:26:51.947 --> 00:26:54.947 For two years, I've done a silver hat. 00:26:54.947 --> 00:26:59.947 See the head tilt movement back and forth is actually, there's three people talking now. 00:26:59.947 --> 00:27:02.947 And by him, Pentao. 00:27:02.947 --> 00:27:04.947 Private security. 00:27:04.947 --> 00:27:06.947 The client wants his legend. 00:27:06.947 --> 00:27:08.947 Some people would like to ask him questions. 00:27:08.947 --> 00:27:11.947 Slowly, the client pays 100% upfront. 00:27:11.947 --> 00:27:12.947 I would like to be paid. 00:27:12.947 --> 00:27:17.707 is also a so in the case all expenses first class tickets this is why we work 00:27:17.707 --> 00:27:22.367 together name of one and called more the top tray is a normal pieces files on a 00:27:22.367 --> 00:27:27.467 laptop sounds like a little describing every room cameras 14 00:27:27.467 --> 00:27:32.687 leaves in it quiet is my second favorite this profession I will not ask what the 00:27:32.687 --> 00:27:39.387 first is the phone rings when the helicopter is ready 40 00:27:39.387 --> 00:27:45.927 cutscene over thank goodness click to take control all right aside from the 00:27:45.927 --> 00:27:56.147 sounds being just totally messed up this is actually pretty decent okay the cat is 00:27:56.147 --> 00:28:03.187 there's definitely some geometric issues what are we supposed to do take the 00:28:03.187 --> 00:28:10.867 briefcase when you're ready okay where is it well pet buttons don't look at me 00:28:10.867 --> 00:28:17.967 like that buttons Easter egg man's gotta eat okay where's the briefcase all 00:28:17.967 --> 00:28:24.447 right we have inventory zip ties okay eat it continue all right is it gonna put 00:28:24.447 --> 00:28:30.827 Miami, palm trees, humidity, and one very bad man who preys on people who knit. 00:28:30.827 --> 00:28:31.327 Mmm. 00:28:31.327 --> 00:28:32.327 Champagne. 00:28:32.327 --> 00:28:33.327 Perhaps. 00:28:33.327 --> 00:28:34.327 Coffee. 00:28:34.327 --> 00:28:35.327 Black. 00:28:35.327 --> 00:28:37.827 Like my, well, coffee. 00:28:37.827 --> 00:28:38.827 Black. 00:28:38.827 --> 00:28:40.327 Asset confirmed at the resort. 00:28:40.327 --> 00:28:41.827 Poolside at noon. 00:28:41.827 --> 00:28:42.827 Oh, poolside? 00:28:42.827 --> 00:28:44.827 Housekeeping enters the suite at 10 and 2. 00:28:44.827 --> 00:28:45.827 Camera coverage. 00:28:45.827 --> 00:28:46.827 Lobby. 00:28:46.827 --> 00:28:47.827 Elevator. 00:28:47.827 --> 00:28:48.827 There is no dog. 00:28:48.827 --> 00:28:49.827 Blind spot. 00:28:49.827 --> 00:28:51.567 The west service stairwell. 00:28:51.567 --> 00:28:55.007 I think we just saw a glimpse of the pool. 00:28:56.447 --> 00:28:56.707 Alright. 00:28:57.747 --> 00:28:58.867 Oh, great. 00:29:00.067 --> 00:29:01.027 Azure Palm Resort. 00:29:01.207 --> 00:29:02.187 Click to take control. 00:29:03.387 --> 00:29:08.667 It's even got like the like low rent wallpaper that you'd expect at one of these joints. 00:29:09.227 --> 00:29:09.867 Oh, wow. 00:29:10.487 --> 00:29:11.947 This is actually like kind of intricate. 00:29:12.107 --> 00:29:12.407 Oh, good. 00:29:12.627 --> 00:29:12.867 Okay. 00:29:12.867 --> 00:29:15.307 I was worried we wouldn't be able to just walk through the walls. 00:29:15.607 --> 00:29:16.647 But okay. 00:29:16.647 --> 00:29:18.087 We can toss a coin in the fountain. 00:29:18.487 --> 00:29:18.987 Your wish. 00:29:19.067 --> 00:29:20.127 No more CMOS errors. 00:29:20.227 --> 00:29:20.827 50-50. 00:29:21.567 --> 00:29:27.687 It seems like Steve's shop in the airplane scene is like oddly placed inside of the location 00:29:27.687 --> 00:29:28.987 we're supposed to be at. 00:29:29.287 --> 00:29:30.387 Let's just ignore that. 00:29:31.647 --> 00:29:35.427 So check in at the front desk and then reach the 17th floor. 00:29:37.427 --> 00:29:39.227 Welcome to the Azure Palm. 00:29:39.787 --> 00:29:41.947 Checking in, Mr. Me. 00:29:42.667 --> 00:29:43.227 Coulter. 00:29:43.667 --> 00:29:44.787 Stephen Coulter. 00:29:45.027 --> 00:29:45.607 Mr. Me. 00:29:45.647 --> 00:29:47.047 I believe I have a reservation. 00:29:47.767 --> 00:29:50.867 Room 1408, partial ocean view. 00:29:51.567 --> 00:29:52.407 I do apologize. 00:29:52.827 --> 00:29:54.667 The ice machine on your floor is out of order. 00:29:55.327 --> 00:29:56.567 Story of my life. 00:29:59.907 --> 00:30:02.007 All right, we now have a room key in our inventory, 00:30:02.247 --> 00:30:03.707 so is this the elevator? 00:30:03.707 --> 00:30:04.107 Oh. 00:30:08.987 --> 00:30:09.927 What is that? 00:30:11.747 --> 00:30:12.267 Reach. 00:30:12.507 --> 00:30:14.247 Okay, west service stairwell. 00:30:15.127 --> 00:30:16.847 Let's just assume this is north. 00:30:20.427 --> 00:30:20.947 Good. 00:30:21.567 --> 00:30:51.207 like what well that is frustrating because it gave us this really like 00:30:51.207 --> 00:30:55.687 anticipatory music. I'm sure there's a far better terminology for what I was about to say. 00:30:55.847 --> 00:31:02.227 My name is Steve. Oh, no, don't. I'm not. Okay. Yep. Most of my clients are old enough to remember 00:31:02.227 --> 00:31:09.467 when a hard drive was a luxury. All right. We got most of the way, I think. This is a test that 00:31:09.467 --> 00:31:15.527 more often than not, we don't complete it. I will say the elements of this judged individually. 00:31:15.967 --> 00:31:19.367 Definitely the pieces were all there. I don't know that the puzzle made the correct picture 00:31:19.367 --> 00:31:24.747 that was shown on like the puzzle box but it was pretty darn close and I saw some truthfully this 00:31:24.747 --> 00:31:29.187 did a better job than I expected especially considering like the subway scene game and 00:31:29.187 --> 00:31:36.887 things like that so I'm actually I'm okay with this so the final thing was of course the street 00:31:36.887 --> 00:31:42.307 yeet game which is weird because this piece of text moving up and down like that while this one 00:31:42.307 --> 00:31:47.347 remains completely static makes me uncomfortable in some level so this is a Bijan issue 00:31:50.475 --> 00:31:59.535 to just run it in a server. One fist, zero mercy. I already like this. Oh, it seems like 00:31:59.535 --> 00:32:06.575 good music all right all right all right 00:32:06.575 --> 00:32:12.295 let's just explore the city real quick 00:32:12.295 --> 00:32:20.335 i don't even know okay so i saw a car get like 00:32:25.175 --> 00:32:26.875 this one's actually kind of fun 00:32:26.875 --> 00:32:39.595 excuse me excuse me pardon me it does like a like a 00:32:39.595 --> 00:32:43.995 okay i don't know that we hit is that a can we eat the bike 00:32:43.995 --> 00:32:48.115 bonk 00:32:48.115 --> 00:32:56.275 Oh, jaywalking pedestrian. 00:32:59.395 --> 00:33:05.495 Can we punch this parking meter? Yeah, we can just everything except the buildings. All right. 00:33:05.495 --> 00:33:09.495 There's unfortunately the orientation of some of these things is not quite right. 00:33:09.755 --> 00:33:12.475 We can see that we're moving sideways when we move forward. 00:33:12.475 --> 00:33:19.435 and the cars are also sorry pal 00:33:19.435 --> 00:33:28.275 and it's interesting how they uh interact with the scenery as well can we eat this statue i don't 00:33:28.275 --> 00:33:37.055 think we can but oh yeah we can all right i will say the cartoonness of this game is actually 00:33:37.055 --> 00:33:42.175 kind of well done i don't know how easy that is to assess just watching someone play it but like 00:33:42.175 --> 00:33:43.895 The pacing is decent. 00:33:44.035 --> 00:33:46.795 I think maybe that's attributed to the music as well. 00:33:47.135 --> 00:33:50.635 But the way that when you actually, all right, the loop is getting a little repetitive. 00:33:50.955 --> 00:33:56.135 When you actually yeet the things and they also go and interact with other items in the environment. 00:33:56.135 --> 00:34:01.335 So like how that will bounce off the building and then maybe hit other things in the path. 00:34:01.575 --> 00:34:02.535 That's pretty cool. 00:34:02.655 --> 00:34:08.015 And this was, this was actually quite arcade style and playable. 00:34:09.115 --> 00:34:10.495 Our score is going up. 00:34:10.555 --> 00:34:11.835 We have combo multipliers. 00:34:12.175 --> 00:34:17.635 Overall, Street Yeet here is a success. Alright, so that is probably going to conclude our first 00:34:17.635 --> 00:34:23.955 look and test of the general availability DeepSeek V4 Pro, the 0813 model designation. Now, this is 00:34:23.955 --> 00:34:28.795 definitely an improvement over the preview version. That felt pretty undercooked, and I was not at all 00:34:28.795 --> 00:34:33.435 impressed with it, but if we look back at some of the tests that were run in that video, which is on 00:34:33.435 --> 00:34:38.795 the channel, we ran some of the same ones today. This is a significant improvement. I would say, 00:34:38.795 --> 00:34:43.335 though, I think the Flash model impressed me more than this, especially considering the size of it. 00:34:43.655 --> 00:34:48.895 This has a bit more competition, but again, it is a big leap on its predecessor, which I guess is 00:34:48.895 --> 00:34:53.615 what we like to see. I was impressed a few times with some things, like basically results within 00:34:53.615 --> 00:34:58.935 results. The cinematic Steve the PC repair game, it actually had some elements to it that were like, 00:34:58.995 --> 00:35:02.715 hey, you know, it kind of thought this through. Just the way when they were speaking, their heads 00:35:02.715 --> 00:35:06.795 were like moving back and forth, not the weird chin thing that was happening. I'm not sure what 00:35:06.795 --> 00:35:11.335 happened there. The music it included in both that game and then the Street Eat game were actually 00:35:11.335 --> 00:35:16.535 like they definitely fit the scene that it was drawing here and it did a good job making the 00:35:16.535 --> 00:35:21.835 Street Eat game very cartoonish and the playability of it felt good I guess for lack of a better term. 00:35:22.335 --> 00:35:28.375 The model that we got of the engine was actually it wasn't printable so unfortunately maybe it 00:35:28.375 --> 00:35:33.315 failed to like check that box in the specific task however if we just look at one of the previews 00:35:33.315 --> 00:35:38.495 right here. It did a pretty decent job, definitely better than the other models I've tested this with, 00:35:38.755 --> 00:35:43.255 though I have not tested this specific prompt with any big beefy Frontier models yet, so 00:35:43.255 --> 00:35:47.715 keep that in mind as well, but it definitely had some chops here, and the amount of research it did 00:35:47.715 --> 00:35:53.055 to actually get to that point was quite a lot. We have all the things in here, downloading these 00:35:53.055 --> 00:35:57.115 images when it doesn't have vision, so I'm not quite sure what to make of that, but nonetheless, 00:35:57.115 --> 00:36:03.615 I think probably the biggest point of frustration for me is it just goes all out when trying to 00:36:03.615 --> 00:36:09.455 actually ensure that every aspect of every line of code it made is correct. And with a model like 00:36:09.455 --> 00:36:15.255 this where it does not have vision, sometimes it takes longer and there is a cost thing to it as 00:36:15.255 --> 00:36:20.415 well. Assuming you're going to be paying to use these models, it just sometimes is simpler to 00:36:20.415 --> 00:36:25.015 have it make the result, not try to do like overviews and checks and follow-ups. You look 00:36:25.015 --> 00:36:29.795 at it yourself and then tell it what's wrong and tell it to fix it. Having to just go all out and 00:36:29.795 --> 00:36:35.955 it trying to fix it adds a lot of time and expense and tokens to the task, which maybe is more of a 00:36:35.955 --> 00:36:41.115 consideration when doing zero shot tests like this. But I find that something I've consistently had 00:36:41.115 --> 00:36:46.555 frustrations with as well as with the flash model, where if I'm having them make like a street heat, 00:36:46.615 --> 00:36:50.455 it's easier if they just spit the script out, I test it and say, hey, the buildings are disappeared 00:36:50.455 --> 00:36:54.915 and then it fixes it. Rather than it trying to go through and like, well, let me try to spin up a 00:36:54.915 --> 00:36:59.455 headless browser and then I'll take a screenshot and like do a heat map because it can't actually 00:36:59.455 --> 00:37:06.135 see it or some other function. So it's a little token happy. I did run it on the highest potential 00:37:06.135 --> 00:37:12.095 reasoning mode so that could be attributed to it as well but I still would like to be able to do 00:37:12.095 --> 00:37:16.655 that without it like going all out to try to make sure it's polishing every aspect of the script 00:37:16.655 --> 00:37:21.615 because I want max performance. I just don't need that level of depth. So that may be a personal 00:37:21.615 --> 00:37:26.335 complaint. But you know, so for the final thing, let's take a look at everything we spent for the 00:37:26.335 --> 00:37:33.375 entirety of this test. We can see right here that all cost $1.91. So that was pretty darn cheap. 00:37:33.435 --> 00:37:38.715 And you do have to remember this is going to be increased their API prices. They say significant 00:37:38.715 --> 00:37:43.755 increase expected, but I mean, a significant increase might be like, okay, the entirety of 00:37:43.755 --> 00:37:49.875 this test was $8 or something, which is still in perspective, not horrible. The Quinn 3.8 max test 00:37:49.875 --> 00:37:55.775 video that we did a few weeks ago or like a week or two ago, that was about $32. So keep that in 00:37:55.775 --> 00:38:00.415 mind as well. That may be a bit more performant than this, but there is going to be a pricing 00:38:00.415 --> 00:38:06.435 consideration as well. And this was very, very, very cheap. It won't remain that cheap, but even 00:38:06.435 --> 00:38:11.035 if it gets five times as expensive, it's still not horrible. I suppose that's going to conclude 00:38:11.035 --> 00:38:16.975 our first look and test of the 0813 DeepSeek V4 Pro model, the general availability version. And 00:38:16.975 --> 00:38:22.615 this is now out of preview and that's just pretty exciting. Basically now I think I'm going to test 00:38:22.615 --> 00:38:29.135 Grok 4.6 because that also came out today and then I believe Friday QN 3.8 27b is coming out so that 00:38:29.135 --> 00:38:33.535 will be exciting as well. Seems like there's been this has been like a pretty busy week in terms of 00:38:33.535 --> 00:38:38.555 auto releases so that's cool also and when MuseSpark 1.2 comes out that's going to be awesome. 00:38:38.955 --> 00:38:43.675 I don't know how big it is but I'm very excited to try to run it locally. So with that if you have 00:38:43.675 --> 00:38:47.235 any questions please feel free to leave them in the comments and thanks for watching.