start	end	text
0	2260	shooting out. I believe that was something I've seen.
5200	11240	The trees are quite reminiscent of a poo. Deep Seek V4 Pro has just been released in
11240	15620	general available version, which also happened with the V4 Flash model back
15620	22260	about 12 or 13 days ago on 0731. Now that model, when tested, proved to be a significant,
22460	26900	significant increase in capability over the preview version that was previously around.
26900	31620	and today we're going to be finding out if the same can be said for the newly updated Checkpoint
31620	37900	general availability for V4 Pro dubbed 0813. So before we get into it do feel free to subscribe
37900	43240	so we can hit 70k and then 100k with the plaque and with that basically there's not a lot to say
43240	47500	about this right now because it seems to have been more or less silently released. There's no
47500	52000	official announcement post yet on X which I do believe is generally where we would see this
52000	56140	but they do mention here just in the website a couple references to it where
56140	63500	If you use this model identifier right here for V4 Pro, you are being routed to the new V4 Pro 0813 version.
63820	72000	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.
72160	79240	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,
79240	85900	we can see that the model has a million context window with a maximum output length of 384k, which really is a lot.
86140	91520	Now, something that can sometimes happen with DeepSeek, at least historically, is there may be a lot of thinking.
91940	98140	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.
98140	105360	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.
105820	113760	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.
113760	118740	There were some references and photos floating around which apparently came from a group related
118740	124000	to DeepSeek. I believe they said it was a WeChat group about DeepSeek or something. And if this is
124000	128160	to be believed, we can see that these would hypothetically be the benchmark results for
128160	134600	DeepSeek V4 Pro 0813 or the general availability model. And basically my main interest here is
134600	140080	just seeing the leaping capability over the V4 Pro preview to now because I was initially
140080	146160	disappointed with the v4 pro preview as well as the v4 flash preview they felt very undercooked and
146160	151200	probably because they were named preview that's actually like fair it makes sense but again i
151200	156340	want to reiterate the leaping capability between flash and then the 0731 general availability
156340	162300	version of flash was absolutely massive this is like the best local model right now for like a
162300	169200	dual spark setup so i mean in a niche you know but still if that leaping capability remains then what
169200	175660	we can see right here, just in this deep SWE score increase from 12.8 to 62.7 would then
175660	181040	hypothetically put this above things like Opus 4.8 and make it very, very competitive. Being
181040	186220	especially that this is an open weight model or will be, that's also extremely exciting because
186220	192700	this is not like horrifically gigantic. I guess when considering things like Kimi K3 or Quen 3.8
192700	198620	max, it's still big at 1.6 trillion parameters. It is a mixture of experts model with how many
198620	204060	active. So if we do estimate that this is going to be the same size as the preview version of V4 Pro,
204180	210080	it's a 1.6 trillion parameter mixture of experts model with 49 billion active. So it is still very,
210200	215640	very large, but it's not as gigantic as like the closing in on 3 trillion parameter models. So
215640	220800	additionally to that pricing, and this is something that we need to basically take just as it is right
220800	225120	now, because they mentioned and have been for quite a bit of time that their API pricing is
225120	230280	going to be significantly increased soon. Now significantly is relative to what it currently is,
230340	236220	which is like very, very, very cheap. So right now pricing per million input tokens for DeepSeek
236220	244800	V4 Pro is 4.35 cents. So, okay, now I know what this is, but like last time I messed up on this
244800	251720	one and I said it was two cents and it wasn't, and then I got roasted. So this is 43 cents per
251720	257060	million input tokens and 87 cents per million output tokens is the current price. Now, if this
257060	262360	gets like a significant increase, say it goes up like five or 10 times, it would still hypothetically
262360	267480	be competitive. So it's all relative to what it currently costs. And with that, I am going to be
267480	273700	using this from within Codex. When we tested V4 Flash, the general availability model on the 31st
273700	278640	of July, I did also use it through Codex because I believe they had suggested to do so. And the
278640	284280	performance was very very good I was happy with it so I'm going to be using Codex as well for this
284280	292060	now I should also mention that they did release in beta a desktop app for Codex or ChatGPT so I
292060	296220	also have it routed and working through this desktop app as well and it is doing a skate game
296220	300540	for us in C++ I would imagine that will take quite a bit of time so I did just want to start
300540	304840	it and verify this app was actually working properly on this system and it does seem to so
304840	311400	keep in mind any mention of GPT or codex here is routed directly to DeepSeq on max reasoning effort
311400	317580	for the 0813 model. And with that, let's take a look at our, okay, so apparently the OpenAI
317580	324860	beta application set itself to be the default for all file types. That's infuriating. All right.
325160	331540	All right. Very Quen 3.8 aesthetic. I'm going to turn sound on because I did see some mentions of
331540	338340	it just talking about like sounds. All right. It starts in our notepad. Can we resize this?
338680	345560	All right. Acceptable. Minimize. Unminimize. Good. The big one. Is there a right click? Yeah,
345620	352660	there is. There is also the correct date and time in my locale in the bottom right. And there is
352660	356840	what I would imagine here to probably be the special feature. It is the holographic desktop
356840	362180	feature so we will save that for last. Let's start out by looking at the files app which I don't
362180	366480	think I've seen a files icon that looks like this before. All right we have some nice transparency
366480	371680	to the application backgrounds and definitely a nice synthwave theme which I do personally like.
371820	375520	So unfortunately it seems our fake file system is just static and we can't browse through it.
375520	380240	Nonetheless can we create new things in it? Okay we can. Can we remove the new things we created?
380240	387660	we also can can we remove things that were there by default yeah we can all right cool that makes
387660	393600	me a little happier and then up would just go there oh okay not bad decent bit of immersion
393600	398660	at least in terms of what directories would exist were this a real computer notepad we did kind of
398660	402880	see because it opened with this just to give us some information the only thing i like to see is
402880	408320	will it save it as a text file on the system no such directory didn't i delete the documents
408320	413700	directory. Hold on a second. That's okay. Oh, it remembered that I did that. I might have messed
413700	420820	that up. So that's going to be, that's on me. Okay. This right here is just not right. It should not
420820	427060	have a like visible text box that happens. Sometimes help did give us things. Let's just
427060	437640	do Neo fetch. That's actually pretty well done right there. Negula or nebula, nebula,
437640	441700	still, that's actually alright. I'm just being kind of mean.
442220	443880	Memory, this is fine.
444880	447640	Oh, it actually put a bit of cleverness in there.
448240	450760	1.44 megs out of infinity, this is fine.
451080	453300	Isn't that how much an old school floppy disk would hold?
453840	455560	I don't know, because those are older than me.
456020	457580	Next up, calculator.
457580	461740	Okay, so 47 times 8.
464760	465200	376.
465620	467040	I haven't done one of these in a while.
467040	469980	divided by 6. Alright. Settings.
472400	475120	Definitely seeing a bit of like Quen 3.8
475120	478680	in here just in terms of some of the style here. What is that?
478800	479780	And why do I like it?
481505	490685	aurora oh okay sunset city this no one would understand that reference this looks like
490685	495125	something one of my friends in high school was making as a game project it actually looks like
495125	503125	a ton like the splash screen for his project we have customizable accents as well difficulty easy
503125	509205	normal and hard all for games okay i'll probably set it to easy as you may be familiar with my
509205	517045	quick to anger game playing style next up let's try our games the speaker is on so city heist
520885	522405	oh you absolute
525125	532305	Okay, so you have to press enter to start and not click the button.
532305	533305	Please shift gears.
533305	536305	Oh, okay.
536305	537305	Okay.
537305	542445	All right, it's...
542445	547445	Oh, cool.
547445	554125	Oh, okay.
554125	564805	troubled but the sounds for like this smack are kind of funny like all right
564805	569165	and we wrecked and we only grabbed a hundred out of a thousand but it did
569165	572845	work so keep in mind we need to press enter to play these games and not hit
572845	588105	the button that says launch. Oh, all right. Do I go through these? Yeah. I just assume based on
588105	592505	the sounds. I wonder if that's how like dogs are since they don't understand language, but like
592505	596465	that to me seemed like a positive reinforcement sound. So it makes me want to do it more.
596465	599905	And I think we did just discover something about how dog screens work.
599905	608285	okay the score I don't know that we need that many decimal points just to show a score I think
608285	615025	maybe just rounding up to 734 or down would have been acceptable but I applaud it for its
615025	620825	intricacy and then oh read me okay so our special feature is something like holographic
620825	625665	and I would imagine having a couple apps open just when we trigger that is probably the move so
625665	631085	oh yeah all right so it just puts like an interesting card effect this is what it looks
631085	636245	like this looks like an apple vision pro like app demo this is the way they look from inside the
636245	644805	vision pro headset just if you have used one you may resonate with what i'm saying okay overall
644805	650125	it was fine everything worked nothing really blew me away but again that may be a limitation of the
650125	656165	test and less of a limitation of the model. So alas, it's kind of neat. We will continue.
656405	662125	So our skateboard game is still going. And again, I'm using this with the newly released beta version
662125	667005	of the chat GPT desktop app. I would imagine this will take quite a bit of time. So let's check to
667005	671725	see just like, oh no, because we already have looked at the cost and keep in mind, this is
671725	678585	partially for the skate game as well as the browser OS. So 20 cents total across two separate tasks,
678585	685685	one which is still ongoing all right so after 68 minutes and 24 seconds skate nyc is complete
685685	691145	so oh you idiot you spoiled it with these screenshots and you don't even have vision
691145	698925	i'm infuriated nonetheless now that we know what the game looks like i had to stop myself
698925	704785	from saying something a little more colorful let's just take a peek at the game all right so
704785	715885	here's our NYC skate game should be in outputs right yes all right it's not quite as good as
715885	721865	the quen 3.8 model but that model does have 800 billion more parameters
721865	729385	does that say NYC skate on the ground it's inverted but it does and it says established
729385	738865	1999 all right sounded like the beginning of the Van Halen song I don't know what exactly is all
738865	744745	right let's just maybe not do that same thing there are sound effects though I don't normally
744745	752305	hear sound effects can we actually move beyond all right not quite I will say it's not wonderful
752305	759325	that's actually kind of cool though the way it like stylized the game title nice nice
759325	766625	well we don't need to do the drum kit sim the trees are
766625	775765	quite reminiscent of a poo if i were to be honest but all right the legs are like kicking
775765	786545	this may be the funniest skate game result we've ever seen which is you know something all right
786545	793045	we have a street sign there the text is a little sideways as are the building faces this is
793045	798185	definitely a troubled result however i do like the water fountain shooting out i believe that
798185	807165	was something I've seen. I've seen that in, I think the Quinn result. Now here's what I would
807165	812365	like to just quickly do for some reference. This is the result we received for the C++ skateboard
812365	818825	test from the preview version of DeepSeek V4. We can see right here, I'm going to turn the speaker
818825	825505	off. This is pretty bad. It was like so bad that I actually remember to go to this specific thing
825505	831885	just to showcase it. So the leap in capability based off of what we see right here to what we
831885	836865	just saw in the skate game that it had created for us, which we can just look at in these screenshots
836865	842405	here, is still pretty significant. Next up, I'm giving it the subway FPS test. I've also added
842405	846625	on to the end of the prompt. Don't worry about checking over the script. I will manually check
846625	850645	it and if there are issues, I will report them back to you just because this will really like
850645	857225	extend every single result by like two times in terms of like time-wise length just in the interest
857225	861345	of something I could do visually in like five seconds and then be like hey this doesn't work
861345	866605	so we're gonna do that just to maybe like keep things a little tighter all right so it did finish
866605	871385	the subway fps game the problem is it just keeps trying to do like troubleshooting which
871385	876705	it's not good at doing on this system for some reason all right let's check out our subway fps
876705	884185	Montrand Avenue. All right. I'm going to turn the speaker on. So it gave us two options. We can
884185	893185	either explore the station. Okay. It's, Hey, you know what? I think that is the best subway car
893185	898485	that I've ever seen drawn in this test. And you're going to be like, well, it doesn't look that great.
898485	903365	I know. It's just that they never really include these. And when they do, they're not really very
903365	910125	good. So this actually, what was that? If that thing starts moving, I would have probably become
910125	916645	more excited than would be reasonable for, you know, responding to such thing. This is very clean.
916765	920925	It's simple. It's nothing that's blowing me away, to be completely honest with you. Queens Boulevard
920925	925905	line. Okay, we have stairs. We have some lighting elements and things of the sort. We have ace,
925905	933665	if you want to go that side. And then tunnels, like indicators for the train's trash bin,
934425	939385	floating particles, more simple, and we can't go up the stairs, which is fine. We don't want to.
939705	946245	Now, of course, the other feature that was included here is the FPS. So let's go back to
946245	951165	menu and then enter the survival FPS. Let's see if it changes the lighting or anything. Oh my,
951165	962925	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
962925	970905	see any enemies look at when we run it's like a waddle all right so uh the big thing I'm
970905	973705	noticing here is there are no zombie enemies so
975887	978507	Good reload. The waddle is just kind of funny.
982827	987247	Weird. Hey, look, there are shell casings that actually drop. That's kind of cool.
987247	992427	All right, this is like, yeah, it's kind of mid. So I'm going to give it a feedback piece.
993187	996727	Now I stopped it because it was just like, it had finished and it was like,
996907	1000487	I can't use this in the headless environment to check. And at that point, it's easier to
1000487	1004227	just check it and tell it what's wrong. So I'll give it that honest feedback.
1004387	1008907	So supposedly our fixed and polished version of the Subway FPS is completed.
1009347	1015347	Let's just take a peek at it. The sound is on. All right. Interesting. I think it swapped the
1015347	1022487	positioning of the buttons. So, ooh, all right. Seven hostiles. It's showing me in the top right.
1022847	1030407	I unfortunately just can't locate them. I would have expected it to have done better. And again,
1030407	1035527	this is one where we could go back and reference the previous model of this. So basically the
1035527	1043347	preview version, it was significantly less detailed than this was. Though unfortunately,
1043347	1048887	we don't have enemies and given the amount of time it took as well as the kind of multiple times I
1048887	1053267	told that there were no enemies I'm regrettably going to call this one. Next up I'm giving this
1053267	1060107	a difficult task to create a 3D model of an RB26 inline-six engine. It must be something that can
1060107	1066227	be 3D printed with little to no support and fit inside of it a tiny DC motor called an N20 motor
1066227	1071347	because this is something that I've told it should like work with my RC car project. Actually I don't
1071347	1075767	think that's specifically in this prompt but that's the whole point of this where this is it
1075767	1080767	should understand and it will scale it properly to that motor so far I've run this test with a few
1080767	1084707	intelligent models and I've never really received anything that was actually worth printing so
1084707	1090287	hopefully sooner than later that will change so our 3d model task is pretty interesting because
1090287	1095207	it's just doing a lot of research on like the actual proper orientation of specific items on
1095207	1100307	this engine like where's the downpipe which is an exhaust component what side is it what is the
1100307	1105927	intake side? What is the exhaust side? Intake, plenum, left or right? So I do like to see the
1105927	1110307	amount of like intricate research that's going into the culmination of this model creation.
1110647	1116427	And it would be really awesome for like us to get a model for this task that's actually printable.
1116507	1121987	It would be the first one. So in one hour and 10 minutes, we received our 3D model response.
1122167	1127647	This is a very, very intricate, just bit of text about the thing it made. Now, of course,
1127647	1132567	the other part of that is actually just taking a look at what it looks like though once I remember
1132567	1137267	what I named this folder we will be able to take a peek assembled display stl I would imagine that
1137267	1146047	to be the case please be good I've really like I need a good one for these actually it's sideways
1146047	1156167	son of a gun this did a really good job this is actually like so far the best one it's not
1156167	1163267	perfect we have some sideways turbocharger issues though i mean this is all right let's let's peek
1163267	1167367	at this in open scad there are a bunch of individual pieces for this so we also want to
1167367	1171307	see if the engine that we told it needs to fit inside the little remote control car engine
1171307	1178147	let's see if it has any hope of properly fitting research it has a research folder right here just
1178147	1185587	showing like a bunch of images and stats for this motor very cool though and here we have some better
1185587	1195667	preview renders. This is definitely well done. Turbo side, okay, underside, all right. It's not
1195667	1200367	something that's going to really easily be 3D printable, although I guess I can't make that
1200367	1205347	determination because this comes in multiple pieces. So it's not designed to be printed as
1205347	1210587	we first saw it just in one lump like this. It's designed to be printed separately. I would like
1210587	1216727	to see. Let's look now at just the single first piece, which is ideally where it would be placing
1216727	1224227	the 3D printed motor that would go in it. Unfortunately, printing this without supports
1224227	1229947	is just not an option in any potential orientation that we see right here, which is a bit frustrating.
1230107	1234187	It almost seems like it has these posts and holes in it to be able to snap fin on other parts,
1234307	1239807	which is cool. Okay. Yep. And I think those are definitely like attachment posts, which is pretty
1239807	1244447	cool. This one would be able to be printed more or less without any supports if we were to
1244447	1251167	put it flat. Then this should have the Nissan logo or at least a text Nissan written on it.
1253107	1259347	Yep and this one's quite well done. No issues here. The oil pan. Oh no. Okay this is where we
1259347	1263967	this is where we decide we're not going to be trying to print this. I'm not disappointed
1263967	1270927	but I'm not super impressed either so I suppose I am lukewarm when it comes to judging this
1270927	1276607	definitely if we just look at this as a render even forget about looking at the individual parts
1276607	1282127	there because it's not really going to be feasible to print this it did a really decent job just
1282127	1286827	assembling it the research that went into this I think is almost probably like I hate to sound
1286827	1291527	like an AI here but like the story isn't the model it's the research that went into it that's
1291527	1298427	generic like AI slob. It really did a lot of decent research here. I mean, did it find this
1298427	1304887	and download it just to see? No. Okay, so this is something it maybe made or maybe probably found
1304887	1309907	online just to be able to better assess what the size of this asset is. But it took a lot of care
1309907	1314947	to ensure that what it tried to generate would have hypothetically fit the specific specifications
1314947	1321347	that were denoted in the original prompt. So not half bad. I'm also going to give this a simple
1321347	1325507	front-end design test where it needs to create the watch website. So it has to have a pretty
1325507	1330987	cinematic panning effect over a 3D watch asset that it creates. I didn't specifically, okay,
1331067	1335787	no, I did specifically denote that the watch needs to be a 3D model. So we'll see how competent it is
1335787	1341627	at putting the entirety of this together. So in one hour and 30 minutes, we got our watch front-end
1341627	1347347	result. It again seems to have a pretty substantial like just list of what was built here. So let's
1347347	1355187	take a look and see how it is. Beautiful watch front end website. That's actually so close to
1355187	1362527	excellent. The big problem is this, I would assume is the watch strap sticking out. This is 3D. So
1362527	1368427	don't be fooled by its 2D-ishness look. It is actually 3D. It's slowly panning too. I really,
1368747	1374567	okay. So a couple of things I do like the wooden background and then this like leather card or pad
1374567	1379067	that this watch is placed on. Additionally to that, the face looks really good. The materials
1379067	1385307	look very good. It's just like an overall, it's a high-end aesthetic, which is exactly what we go
1385307	1391427	for when performing this test. Now, that's actually, that right there looks good. The,
1393267	1399347	let's just sit and appreciate for a minute. Okay, yeah, that, maybe I could have done without that,
1399347	1405147	but it's definitely quite something. And then look at the way that goes away as soon as we
1405147	1412627	start to scroll down. Quiet luxury measured in seconds. Sunburst dial. Okay, it has like a
1412627	1417807	material card here. The sapphire crystal is perhaps something I would have maybe omitted from this
1417807	1423267	section, but the dial looks good. Two watches, one point of view. Please render them. Oh, look at that.
1423367	1427487	That is sick. That really looks pretty good. I love the way they're placed on the table.
1427487	1432527	really the big thing letting it down is just this leather slab is inhibiting our view of these
1432527	1438527	beautiful straps this one is the stainless steel option this one is the gold option and priced
1438527	1445387	accordingly i would imagine yeah although i would say the jump in price of like that is the
1445387	1449787	negligible amount to get the gold one i'd probably go with this one personally oh cool and these ones
1449787	1451907	we can freely rotate so that helps a bit
1454247	1458907	watch hands, but that's okay. I'll forgive it for that. But these are pretty darn cool.
1459487	1466947	These are well drawn, I would say, especially the numeral indicators. Look at the way that
1466947	1471907	they actually have multi-material on them. So they're not just like one static material,
1472087	1477087	it's gold and then black infill or silver and then black infill as well. So that's some attention
1477087	1482887	to detail I do like seeing as well as the kind of brushed material look on the stainless steel and
1482887	1488227	the gold. It's probably weird we're like assessing like a 3D watch, but it just shows like some
1488227	1492287	capability. Okay, that looks pretty good right there. And then we have a clean footer as well.
1492287	1499327	I like this. I'm happy with this. Even though it's like not perfect, it's darn close. So next up,
1499447	1503547	just here from within Codex, and I can't show this full prompt because it has an API key.
1503767	1508567	I'm giving this the full-on cinematic Steve the PC repair game test where it needs to create a
1508567	1514047	cinematic game. It needs to create voice assets for these characters just by using an OpenAI API
1514047	1519307	key that I've given it, and it needs to make a fun game that is enjoyable and playable. I will
1519307	1523567	assess this is probably going to take north of an hour so we're just going to basically leave
1523567	1528807	codex here in the background the beta desktop version of it to complete this task and we'll
1528807	1533487	maybe work on some other things in the meantime so our steve the pc repairman cinematic game
1533487	1539567	completed in basically 65 minutes so a little over an hour all right here's how to play i'll
1539567	1544747	probably just do what it says here not assume that it's already running this in some server for us
1544747	1545967	So, okay.
1546467	1548527	Episode one, the Azure Palm job.
1549147	1549707	Okay.
1550147	1550947	Click to begin.
1551127	1551987	Headphones recommended.
1552227	1552947	We have a speaker.
1554647	1555387	All right.
1557687	1559387	My name is Steve.
1559447	1559987	This is good.
1560167	1561327	I fix computers.
1561587	1561947	Well, okay.
1561947	1566667	Most of my clients are old enough to remember when a hard drive was a luxury item.
1567387	1570347	They pay what they can afford and I make it work.
1571007	1572467	It is honest business.
1574007	1574567	Oh.
1574747	1575567	Thank you, dear.
1576067	1578307	The screen kept telling me the date was 1980.
1579107	1582447	And I thought, well, I have been time-traveling long enough.
1582447	1583107	Fresh battery.
1583827	1584907	Your computer now believes it is 2026.
1584907	1586887	I will bring you oatmeal cookies on Tuesday.
1587007	1587727	Same as the rest of us.
1587727	1589867	And give Mr. Buttons a scratch behind the ears for me.
1589947	1592567	Let me walk you out, ma'am.
1593907	1596467	I'll be right with you, Mr. Thomas.
1596467	1598687	The speech overlays itself.
1598707	1599807	You are a good boy, Stephen.
1600447	1604107	If that quiet man needs a computer fixed, you be patient with him.
1604747	1605987	I like punctual.
1605987	1607347	Door is unlocked.
1607347	1608947	Coffee is warm.
1608947	1609947	A man.
1609947	1610947	What is the job?
1610947	1611947	Victor Kane.
1611947	1614947	For two years, I've done a silver hat.
1614947	1619947	See the head tilt movement back and forth is actually, there's three people talking now.
1619947	1622947	And by him, Pentao.
1622947	1624947	Private security.
1624947	1626947	The client wants his legend.
1626947	1628947	Some people would like to ask him questions.
1628947	1631947	Slowly, the client pays 100% upfront.
1631947	1632947	I would like to be paid.
1632947	1637707	is also a so in the case all expenses first class tickets this is why we work
1637707	1642367	together name of one and called more the top tray is a normal pieces files on a
1642367	1647467	laptop sounds like a little describing every room cameras 14
1647467	1652687	leaves in it quiet is my second favorite this profession I will not ask what the
1652687	1659387	first is the phone rings when the helicopter is ready 40
1659387	1665927	cutscene over thank goodness click to take control all right aside from the
1665927	1676147	sounds being just totally messed up this is actually pretty decent okay the cat is
1676147	1683187	there's definitely some geometric issues what are we supposed to do take the
1683187	1690867	briefcase when you're ready okay where is it well pet buttons don't look at me
1690867	1697967	like that buttons Easter egg man's gotta eat okay where's the briefcase all
1697967	1704447	right we have inventory zip ties okay eat it continue all right is it gonna put
1704447	1710827	Miami, palm trees, humidity, and one very bad man who preys on people who knit.
1710827	1711327	Mmm.
1711327	1712327	Champagne.
1712327	1713327	Perhaps.
1713327	1714327	Coffee.
1714327	1715327	Black.
1715327	1717827	Like my, well, coffee.
1717827	1718827	Black.
1718827	1720327	Asset confirmed at the resort.
1720327	1721827	Poolside at noon.
1721827	1722827	Oh, poolside?
1722827	1724827	Housekeeping enters the suite at 10 and 2.
1724827	1725827	Camera coverage.
1725827	1726827	Lobby.
1726827	1727827	Elevator.
1727827	1728827	There is no dog.
1728827	1729827	Blind spot.
1729827	1731567	The west service stairwell.
1731567	1735007	I think we just saw a glimpse of the pool.
1736447	1736707	Alright.
1737747	1738867	Oh, great.
1740067	1741027	Azure Palm Resort.
1741207	1742187	Click to take control.
1743387	1748667	It's even got like the like low rent wallpaper that you'd expect at one of these joints.
1749227	1749867	Oh, wow.
1750487	1751947	This is actually like kind of intricate.
1752107	1752407	Oh, good.
1752627	1752867	Okay.
1752867	1755307	I was worried we wouldn't be able to just walk through the walls.
1755607	1756647	But okay.
1756647	1758087	We can toss a coin in the fountain.
1758487	1758987	Your wish.
1759067	1760127	No more CMOS errors.
1760227	1760827	50-50.
1761567	1767687	It seems like Steve's shop in the airplane scene is like oddly placed inside of the location
1767687	1768987	we're supposed to be at.
1769287	1770387	Let's just ignore that.
1771647	1775427	So check in at the front desk and then reach the 17th floor.
1777427	1779227	Welcome to the Azure Palm.
1779787	1781947	Checking in, Mr. Me.
1782667	1783227	Coulter.
1783667	1784787	Stephen Coulter.
1785027	1785607	Mr. Me.
1785647	1787047	I believe I have a reservation.
1787767	1790867	Room 1408, partial ocean view.
1791567	1792407	I do apologize.
1792827	1794667	The ice machine on your floor is out of order.
1795327	1796567	Story of my life.
1799907	1802007	All right, we now have a room key in our inventory,
1802247	1803707	so is this the elevator?
1803707	1804107	Oh.
1808987	1809927	What is that?
1811747	1812267	Reach.
1812507	1814247	Okay, west service stairwell.
1815127	1816847	Let's just assume this is north.
1820427	1820947	Good.
1821567	1851207	like what well that is frustrating because it gave us this really like
1851207	1855687	anticipatory music. I'm sure there's a far better terminology for what I was about to say.
1855847	1862227	My name is Steve. Oh, no, don't. I'm not. Okay. Yep. Most of my clients are old enough to remember
1862227	1869467	when a hard drive was a luxury. All right. We got most of the way, I think. This is a test that
1869467	1875527	more often than not, we don't complete it. I will say the elements of this judged individually.
1875967	1879367	Definitely the pieces were all there. I don't know that the puzzle made the correct picture
1879367	1884747	that was shown on like the puzzle box but it was pretty darn close and I saw some truthfully this
1884747	1889187	did a better job than I expected especially considering like the subway scene game and
1889187	1896887	things like that so I'm actually I'm okay with this so the final thing was of course the street
1896887	1902307	yeet game which is weird because this piece of text moving up and down like that while this one
1902307	1907347	remains completely static makes me uncomfortable in some level so this is a Bijan issue
1910475	1919535	to just run it in a server. One fist, zero mercy. I already like this. Oh, it seems like
1919535	1926575	good music all right all right all right
1926575	1932295	let's just explore the city real quick
1932295	1940335	i don't even know okay so i saw a car get like
1945175	1946875	this one's actually kind of fun
1946875	1959595	excuse me excuse me pardon me it does like a like a
1959595	1963995	okay i don't know that we hit is that a can we eat the bike
1963995	1968115	bonk
1968115	1976275	Oh, jaywalking pedestrian.
1979395	1985495	Can we punch this parking meter? Yeah, we can just everything except the buildings. All right.
1985495	1989495	There's unfortunately the orientation of some of these things is not quite right.
1989755	1992475	We can see that we're moving sideways when we move forward.
1992475	1999435	and the cars are also sorry pal
1999435	2008275	and it's interesting how they uh interact with the scenery as well can we eat this statue i don't
2008275	2017055	think we can but oh yeah we can all right i will say the cartoonness of this game is actually
2017055	2022175	kind of well done i don't know how easy that is to assess just watching someone play it but like
2022175	2023895	The pacing is decent.
2024035	2026795	I think maybe that's attributed to the music as well.
2027135	2030635	But the way that when you actually, all right, the loop is getting a little repetitive.
2030955	2036135	When you actually yeet the things and they also go and interact with other items in the environment.
2036135	2041335	So like how that will bounce off the building and then maybe hit other things in the path.
2041575	2042535	That's pretty cool.
2042655	2048015	And this was, this was actually quite arcade style and playable.
2049115	2050495	Our score is going up.
2050555	2051835	We have combo multipliers.
2052175	2057635	Overall, Street Yeet here is a success. Alright, so that is probably going to conclude our first
2057635	2063955	look and test of the general availability DeepSeek V4 Pro, the 0813 model designation. Now, this is
2063955	2068795	definitely an improvement over the preview version. That felt pretty undercooked, and I was not at all
2068795	2073435	impressed with it, but if we look back at some of the tests that were run in that video, which is on
2073435	2078795	the channel, we ran some of the same ones today. This is a significant improvement. I would say,
2078795	2083335	though, I think the Flash model impressed me more than this, especially considering the size of it.
2083655	2088895	This has a bit more competition, but again, it is a big leap on its predecessor, which I guess is
2088895	2093615	what we like to see. I was impressed a few times with some things, like basically results within
2093615	2098935	results. The cinematic Steve the PC repair game, it actually had some elements to it that were like,
2098995	2102715	hey, you know, it kind of thought this through. Just the way when they were speaking, their heads
2102715	2106795	were like moving back and forth, not the weird chin thing that was happening. I'm not sure what
2106795	2111335	happened there. The music it included in both that game and then the Street Eat game were actually
2111335	2116535	like they definitely fit the scene that it was drawing here and it did a good job making the
2116535	2121835	Street Eat game very cartoonish and the playability of it felt good I guess for lack of a better term.
2122335	2128375	The model that we got of the engine was actually it wasn't printable so unfortunately maybe it
2128375	2133315	failed to like check that box in the specific task however if we just look at one of the previews
2133315	2138495	right here. It did a pretty decent job, definitely better than the other models I've tested this with,
2138755	2143255	though I have not tested this specific prompt with any big beefy Frontier models yet, so
2143255	2147715	keep that in mind as well, but it definitely had some chops here, and the amount of research it did
2147715	2153055	to actually get to that point was quite a lot. We have all the things in here, downloading these
2153055	2157115	images when it doesn't have vision, so I'm not quite sure what to make of that, but nonetheless,
2157115	2163615	I think probably the biggest point of frustration for me is it just goes all out when trying to
2163615	2169455	actually ensure that every aspect of every line of code it made is correct. And with a model like
2169455	2175255	this where it does not have vision, sometimes it takes longer and there is a cost thing to it as
2175255	2180415	well. Assuming you're going to be paying to use these models, it just sometimes is simpler to
2180415	2185015	have it make the result, not try to do like overviews and checks and follow-ups. You look
2185015	2189795	at it yourself and then tell it what's wrong and tell it to fix it. Having to just go all out and
2189795	2195955	it trying to fix it adds a lot of time and expense and tokens to the task, which maybe is more of a
2195955	2201115	consideration when doing zero shot tests like this. But I find that something I've consistently had
2201115	2206555	frustrations with as well as with the flash model, where if I'm having them make like a street heat,
2206615	2210455	it's easier if they just spit the script out, I test it and say, hey, the buildings are disappeared
2210455	2214915	and then it fixes it. Rather than it trying to go through and like, well, let me try to spin up a
2214915	2219455	headless browser and then I'll take a screenshot and like do a heat map because it can't actually
2219455	2226135	see it or some other function. So it's a little token happy. I did run it on the highest potential
2226135	2232095	reasoning mode so that could be attributed to it as well but I still would like to be able to do
2232095	2236655	that without it like going all out to try to make sure it's polishing every aspect of the script
2236655	2241615	because I want max performance. I just don't need that level of depth. So that may be a personal
2241615	2246335	complaint. But you know, so for the final thing, let's take a look at everything we spent for the
2246335	2253375	entirety of this test. We can see right here that all cost $1.91. So that was pretty darn cheap.
2253435	2258715	And you do have to remember this is going to be increased their API prices. They say significant
2258715	2263755	increase expected, but I mean, a significant increase might be like, okay, the entirety of
2263755	2269875	this test was $8 or something, which is still in perspective, not horrible. The Quinn 3.8 max test
2269875	2275775	video that we did a few weeks ago or like a week or two ago, that was about $32. So keep that in
2275775	2280415	mind as well. That may be a bit more performant than this, but there is going to be a pricing
2280415	2286435	consideration as well. And this was very, very, very cheap. It won't remain that cheap, but even
2286435	2291035	if it gets five times as expensive, it's still not horrible. I suppose that's going to conclude
2291035	2296975	our first look and test of the 0813 DeepSeek V4 Pro model, the general availability version. And
2296975	2302615	this is now out of preview and that's just pretty exciting. Basically now I think I'm going to test
2302615	2309135	Grok 4.6 because that also came out today and then I believe Friday QN 3.8 27b is coming out so that
2309135	2313535	will be exciting as well. Seems like there's been this has been like a pretty busy week in terms of
2313535	2318555	auto releases so that's cool also and when MuseSpark 1.2 comes out that's going to be awesome.
2318955	2323675	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
2323675	2327235	any questions please feel free to leave them in the comments and thanks for watching.