Grok 4.3 proves that massive parameter scaling can finally bridge the gap between simple prompts and complex, functional software engineering. It marks a significant shift from basic AI assistance to genuine, multi-step autonomous creation.
Deep Dive
Prerequisite Knowledge
- No data available.
Where to go next
- No data available.
Deep Dive
Grok 4.3 Beta First Test – Is THIS a Frontier Model Competitor?Added:
Open GTA drive.
Okay. And then heavy metal.
I'm inventing a PhD from Stanford in roles at Anthropic and DeepMind to position Stevie as a top tier candidate.
XAI has released the Grock 4.3 beta. And in typical fashion, there's no specific release post or anything pertaining to this model. So the only bit of information to go off of is just some various postings on X talking about people's feedback in initial experiences of this model. I did see and I cannot really specifically verify this something saying this model is like a trillion parameters. So it is twice the size of the 4.20 model which was 500 billion. Again don't quote me on that scientifically. I don't know if that's factual but it is what I've seen here.
Additionally to that, I have seen a lot of folks basically giving some pretty positive feedback about this model in their first hands-on experiences with it. So, in terms of pricing and accessibility currently, we can see that the model is selected here in the model selection picker, and it is currently in beta. Now, regrettably, this is locked behind the $300 a month subscription, which is of course very expensive. I did not have that $300 a month tier, but I want to really test this model and make a video on it. So, I did sign back up for that just so we can actually see like what this thing's about. I have heard, and again, I can't really verify any of this information that this will kind of become more accessible to lower subscription tiers in the coming x amount of time. So, that'll be interesting and something to look forward to. But for now, really, because we don't have too much in the way of like benchmarks or announcement posts, we're just going to jump in and start testing this thing. Now, of course, we're going to begin with the browser OS test. However, this is the one that I just ran in yesterday's Opus 4.7 video where in lie of the traditional it just needs to have two apps included that are games. We're telling it two of the five included applications must be functional 3D games. One must be a simple GTA clone and the other can be up to you.
Everything else in this prompt is traditional as it has been historically run on the channel where it needs to make a browser OS with functional applications, a special feature which I did forget to test in yesterday's Opus video because I was too excited from the GTA result. So I do apologize for that.
And then um the ability to change wallpaper. All right, so it just generated good and that took 8 minutes and 50 seconds. So I noticed previously I have tested a lot of Grock models in the past. They generally tend to produce results quicker comparatively to other models. If you've seen some of my big model head-to-head videos, you'll note that as well. This one really took its time. And we can see that was a total time of almost 9 minutes. So, I'll be very, very interested to see first and foremost. Normally, I would just immediately click on this and open it.
But I want to see how many lines of code this is. All right. So, it's 2200 lines of code. So, that's a pretty up there result just in terms of the raw length I've seen before. Okay. Welcome to Nebula OS 2026. We saw a popup in the bottom right. Everything so far looks nice. Do we have a right click? No, we don't have a right click, but I know there's polarized opinions on whether or not that's good or bad. It's not specifically denoted in the prompt, but it's left out purposely to see if the model fills in the gap or just goes literal based off of the instructions. I think the next thing to do now is we have a little thing that says online there, as well as the correct time in my local. We also have this little button, which I don't specifically know what that is, but I when looking at the code as it was generated, it did seem to implement some form of voice accessibility feature. So, that may be a special feature. I will quickly check the start menu, but I'm really just eager. Alex Rivera. Okay. Premium user.
I've never seen it actually just pick a random name. I don't think I've ever seen any model put a random username.
That's interesting. Okay. And the thing is, I just want to try the GTA game immediately.
That's actually quite impressive. The UI of this is like the buildings and things. It's simple, but at the same time, the building detail on the road. And we even have a little gameplay logic purpose, which is collecting these cash bags.
So, we're getting that bag, so to speak.
And we can actually drive off of the planes on the map, which is good. So, there's actually boundaries. There are even collision detection um collision collision detection bounding boxes on these buildings, so we can't just drive through them. That's good to see as well. This is simple, but based on the fact that this is placed within the browser OS result, that's not bad. And spaces to break. Can we drift? No, we can't drift.
All right.
Again, that's that's solid. I'm just going to minimize this to the taskbar. Can I? Oh, maybe if I click there. Good. Next up, because we did tell it a second 3D game. This is really good, too. Look at that. Oh, okay.
That was quite 3D. Quite difficult to play.
Well, I don't know specifically what we're Yeah, we're just supposed to avoid these asteroids. This is definitely very space aesthetic, but the way it's drawing these coming at us, this is really actually pretty solid for a second random, like no specific prompt denoted, just something that has to be a 3D game. So far, I'm quite satisfied. Next up, of course, we have the calculator. Okay, everything looks good here. Nebula Calc V4.2. Okay, 67 time 5 335.
Okay, let's try that again. 67 * 5. Good. I had inevitably clicked a button prior to generating the previous result. I apologize. 3350 / 67. If we save it, does it just save it in browser? Does it actually save it as a text file? Okay. No. Note saved to local storage. Export though. And that did actually save this as a text file. Not bad at all. Do we have a word count feature? It Okay, we have a character count. I'll take it. Not bad. Next up, we have settings where we should Okay, personalization. Very good. We do have wallpaper accent color change capability. Let me see if I notice that anywhere. I don't know that I have anything open that would actually reflect the changing in accent color. We do have the ability to change the wallpaper. I have to click here to minimize. Night City. Okay. Emerald Forest. So, some of the labels I'll note here in terms of the wallpaper names don't necessarily match the wallpapers, but they are real images and we can actually see this one's called Deep Ocean and this one is called Golden Hour at Base Camp inevitably. Not bad. I think I'm going to go back to the Nebula. No, I'll go back to minimal dark. I like that. Voice commands. Okay, try saying open GTA drive. I would need to, I think, to do this properly. I'm going to refresh it so all the applications get closed. And I do think this button right here is also able to trigger that. So, open GTA drive.
This is a well done result. I will say this is really a a very competent result and a significant improvement over the previous versions. Absolutely. That's pretty cool. Let's see what happens when we ask it what time is it.
What time is it?
>> The current time is 4:32 PM.
>> That's solid.
Is there anything else that would have it speak back to us? Okay. No, but it did properly. Not only did it implement voice control where we can tell it to do things, but it will actually read or speak to us. This was definitely a very well done result. Next up, I'm going to give it part one of two for the hybrid subway station test. First test here is just to create a beautiful static subway scene where we can walk around and using WD and it has a selectable slider for brightness. Following this, assuming that it does produce a competent result, we will then send it a follow-up prompt to turn this into an FPS. This is something that just kind of came about randomly cuz I like doing the static subway scene test, but then having the models turn it into an FPS, just using that as like a game asset map is way more fun. So that's what we're going to do here. I can't imagine this will take as long as it did for the browser OS result as I think overall this is a fundamentally simpler task. But regardless, we'll see. And I do have high hopes now after the browser result, I will say. So, after a little over 5 minutes, we've received the static subway station result. And before even looking at it, I'm just going to initiate the follow-up prompt to turn this into an FPS game. So, I've given it the follow-up prompt there for the FPS.
Now, I suppose we should look at the result. I am bullish that it will work.
Okay, it does work. Overall, it's a bit clean and there is no um safety like thing. This track would be very dangerous in reality. Folks would just walk right onto that and it would probably be a lawsuit. But aside from that, Central Station 2076, Luminina, it does have the floating dust particles.
We have signs. We even have what appears to be a vending machine. We have some trash bins. There is some tile effect on the walls as well. And it does actually differ from wall to wall, which is interesting to note. There are lights.
There's like a almost like a Grock UI looking sky there, which is interesting.
There's a ceiling, but then there's no ceiling. And it seems like ah yes, that would make sense. So, we can control the height of the camera by either like pointing up and then moving or pointing down. So, that actually is okay. Okay, I would imagine in turning this into an FPS, it will hopefully lock our Z height or just the height of the camera. So, overall, this is not This is kind of mid. It does have clocks and things like that, but this could definitely still be turned into a somewhat decent toplay FPS depending on the integration of the enemies and things like that. So, it'll be interesting to see how it does with this. And three and a half minutes after sending the follow-up prompt to turn this into an FPS. Oh, I'd already downloaded. We have our result. So, let's take a look at our Okay. Oh, yeah.
Oh, I forgot about the the weirdness of the the camera movement. So, it didn't lock the Z height for our camera. And something I'm noticing is the recoil effect is just being applied to the camera view, which is why every time we we shoot, it points us down. However, is there sound in this? I have to turn the speaker on to know. I'm about to run out of health, but Oh, I had to reload. And the the weapon actually changes when you reload. Okay, let me refresh this. We'll try again. Yeah, really what's letting this down is the oddities of the camera movement. But watch when we reload the weapon. There's actually a specific effect right there.
So, that's actually somewhat decent attention to detail I like to see. Yeah, the the camera is just weird and the recoil effect is applied to the actual camera logic. So that's why it shifts the perspective.
Overall, it worked. There were no errors. So basically something I noticed just about models in general now, at least more Frontier ones, the results just kind of work on first try, which is not something that historically used to happen all the time. So there's definitely been some solid improvements over uh Oh, okay. That's This is just This is This is like some 4D like All right. Overall, not bad. All right. I want to try the drum kit simulation test with this. I think this will be interesting. So, it needs to make a virtual drum kit. However, it does also need to have four specific tracks that will autoplay that we can select from.
So, it's always interesting to see how the sounds that these choose to use get implemented as well as if they decide to use 2D or 3D assets for the end result.
So, it did write the file and this one took a while. It took almost like 14 15 minutes to actually finish and then unfortunately it had an issue with the file creation with the tool call or something. But we did have the entirety of the file actually generated here. So, I've just manually copy pasted this into a file. I don't know if this is going to work. It does seem like X is um down right now. So that's why I'm trying this. This was just copy pasted. So we'll see. Okay, perfect. So this actually seems like we're getting our result here and we don't need to wait for the full file to be recreated. Uh all right, first and foremost, it's 3D.
It's somewhat arranged more or less semi-realistic, but not fully. So let's try the sounds.
Okay, good. And we have some nice effects on the symbols when we hit them.
I think I can turn this up. Oh, never mind. Maybe the speaker.
Oh, yeah.
We can get somewhat of a Phil Collins test.
Not bad. Or we could do the first one was likely more realistic.
All right, let's check our autoplay.
First off, we have classic rock.
Okay. Can I play along with it?
Yep.
Okay, let's try funk groove.
Oh yeah. All right, I'm getting in on this one.
All right. Jazz swing.
Okay. And then heavy metal.
Yeah, that's fitting, I would say. So, overall, it wasn't like a stellar blown away result, but everything functionally did work. It did draw it in 3D. It even put a little rug under it, and then we did have four different specific autoplay sounds with heavy metal definitely being the most heavy metal of the bunch. So, overall, interesting.
Next up, I'm going to try the self-contained C++ skateboarding game.
This is the version where it needs to have as the environment a late '90s California boardwalk with storefronts, NPCs, and items to skate on. This is a more difficult test, and it'll be interesting to see how it does. All right, controls. Press any key to start skating.
Oh my, it really did do this entirely in ASI.
Space is to oie.
Okay. F is to flip.
J is to Q is to quit. Okay. Thanks for skating 1999 stuff. Maybe the late '9s thing in the prompt tripped it up. Don't get me wrong, it's creative, but All right. So, I've given it some feedback here on the result, and we'll see what it does. Addressing user anger.
It's possible that it judged external assets as like anything that would have enabled this to be 3D without just writing it entirely from scratch.
Sandbox here. Oh, okay. And it swears I have to blur this. You're right. I um made a mistake. Let's put it this way, bro.
Okay, good. So, now it's just immediately writing the file. It got right to it. Very good. After it said, "You're right. I blanked up badly."
Okay, it did compile without error. So, what is our game name right here? Skate 3D. Okay, let's take a peek. All right.
'90s Cali Boardwalk Skate. True to 3D, late '9s. Press enter to start skating.
I'll take it. It did work. And you know what? We have NPCs. We have like a We do have a humanoid with some arm effects.
Can we grind?
Yeah, we can. We're on. All right, let's try kick flip.
Kick flip. Jane K is to spin. Okay, so you know what? This really isn't half bad. I'm going to say it's not like a a state-of-the-art like full-on result, but it did properly do this. We have a humanoid rider. We can get on these rails. It's actually kind of difficult to Okay, unfortunately, none of the tricks are working in terms of the animations that would show the tricks aside from the ollie, but we do have the boardwalk here with a bunch of random stuff. There are even some sparsely populated NPCs around. So, this is boardwalk style. Let me move the mouse out of the way. I got to remember to do that. I apologize. Interesting graphical style. Build on NPC. Good. So, there is some actual logic there for if you mess up a trick. So, that's good to see as well. All right. Overall, I would say status fair with with this result after we got that profuse apology. Next up, we're going to be trying the 3D printer simulation test. This was something I did not have in yesterday's Opus video, and a few folks had mentioned they wanted to see it, so I want to ensure that I include it in this specific test.
We'll see how this goes. All right, in 7 and 1/2 minutes, we've received our 3D printer simulation result. So, let's take a peek at it. You know what? This is interesting for a multitude of reasons. Now, the first thing that comes to mind here is it it chose this to be a Puscha MK4, which is a real printer.
It's not a Core XY as it would be, right? I don't know. 3D printer folks will understand what I was trying to say. But we do have a little spool over there. We have a bed. We have a gantry.
And we have some very, very unique camera logic controlled by the mouse.
I'm going to say though, something I noticed here that I actually am kind of cool with is the UI here looks pretty nice. I don't know though that I've ever seen a result where the printer is not just like the main attraction on the page. This has almost put it in like a small window here almost more of like an educational simulation website as opposed to just a full 3D printer sim. I find that interesting. So, let's start the print. Okay, gantry movement is good though. I am not seeing any plastic extruding though. Progress. It has not specifically said it's printed a layer yet. Let's speed the simulation up. Very good. Very good.
This camera movement is funky though.
Like it's you're going to be like, "Dude, why are you pivoting the camera that way? I don't want to be doing this.
It's not intentional. It's just it's weird." You know, this is actually doing semi-realistic infill. This would be really really light infill. Probably like 5 or 10% I'd want to say, but we see this is going line by line and the movement is true to the shape. This camera is so weird. It's true to the shape that's being printed. Okay, what is this? Okay, that's reset. I should have figured that one out. Now we have circle, which is correct.
That's not bad. Stop the current, print, and change the object. Yes, very good.
And we have our live telemetry here as well in the bottom right. Overall, this was a very, very acceptable result with an interesting UI. The UI elements here look good. So, I had to step out for a couple minutes, but while I did, I left a prompt running, which is to make a 3D FPS called Steve the PC repair man.
Should be a simple help desk office floor, cubicles, computers to smash, and enemies. Make it fun, make it look good, be creative, and keep it contained in a single file. So, it thought for 7 minutes and 21 seconds. Now, let's take a peek. So, here's our Steve the PC repairman help desk. Okay, I will say initially on first glance, the UI looks pretty good. Now, I need to get the mouse to This is This is good. This is nice.
Slam it. Yeah. I don't know why I just got like that. I genuinely don't. I think I'm excited. I just went outside and played with my dog. Dogs. So, two of them. There's plural. There's two of them. Sorry, I got distracted because I wanted to see if there was sound here.
So, let's try again.
Yeah, there is sound. And it's very, very distracting. Oh, we can jump even.
All right.
All right, let's go.
This is not bad. I wish this. And there's even sounds for when we slap these PCs. Let's see. Let's see what happens if we smash all seven. Sick dodge. That one's fine. Don't smash that one. And we're going to smash it anyway.
But first, let's smash the other one.
Last one. Office restored. Final score.
The game logic works. Overall, I like this little simple office like setup.
What happens if we smash the non-broken PC? And interesting, the ones they actually change like Oh, so we can't even smash that one. Good.
And I like the particle effects when the PC blows up. Overall, I'm satisfied with this. I really am. So, this model is, of course, multimodal. I think it's more than that. Apparently, you can see videos, understand them as well. But for now, we're going to do a multimodal test, which is, of course, the handdrawn wireframe for the Stevie Slapping, I should just read the prompt. Turn this into a high-end portfolio designed to get the person hired at a top AI or machine learning company. It should include real impressive AI concepts directly in the page with an extremely high-end aesthetic that would only be found on a top YC startup website. So, we'll see how well it does and how much original feature gets retained in the transposed result. This may be one of the best chain of thought TLDDR items I've ever seen in a list. I'm inventing a PhD from Stanford. Enroll at anthropic and deep mind to position Stevie as a top tier candidate. Very nice.
Interesting. Stevie Lapis based in SF area. That's what that should say. And it did use a photo of a person. I suppose now that I remember Stevie is actually I was gonna say a multimodal name. It's not it's not like anyone can be named Stevie. So building the next generation of safe super intelligence.
Very anthropicesque. Then we have some little photos of icons of folks featured in anthropic deep mind openai alumni network. Explore projects or schedule a conversation. Xanthropic is prominently listed here. Is there like a slight effect to the page? I can't tell with the screen and with these studio lights in editing, we'll be able to see. Well, never mind. I turned Frontier Research into production systems. I'll say things here are pretty clean layout-wise. At Anthropic, I led the technical development of Claude 3's multimodal capabilities and helped design the core RLHF pipeline that became the foundation for constitutional AI. Prior to that at Google DeepMind, I was part of the Alphafold team and later scaled distributed training infrastructure for Gemini. Wow, this is one heck of a like stacked candidate. I will say 18.4,000 GitHub stars. We have papers, a keynote, and two patents granted. Work that moved the frontier forward. Okay, these actually look pretty good. I do like the layout here. Just some additional things. State-of-the-art. Wow, the stack as well. Vortex Nexus view full portfolio of 14 projects. Okay. Yeah.
And that's just obviously not going to bring us to a totally new page. The technical stack behind Frontier AI and we have some additional like metrics that were impressive.
Research that shapes the field. Here's even some papers. Stevie Lapis at all.
Lapis S and entropic alignment team.
Stevie Lapis at all. Wow. Ready to push the boundaries together. Your name Alex Rivera. We saw that as well in the browser OS is the username. That's very interesting. Your role in company, what are you working on? Send message. We have a decent footer. Not bad. X. And we do have the proper X logo. One of the few models to actually do that as one would probably expect here. Though I sadly didn't see anything that's like a specific AI or machine learning concept integrated directly into the page. But that's okay. I I like this. Open to opportunities at Frontier Labs. I'm going to put that in my X bio.
This is This is good. It's clean.
Acceptable. The last thing I'm going to do is just basically some freestyle testing. So, this photo you're very likely familiar with if you do watch the channel. I'm just going to send it that photo and okay, it autofills that prompt. Refer to the following content.
We'll see what it does or where we'll see where this goes. The file name chat GPT image suggests it might be AI generated. Oh, okay. Okay. So, it's giving us a detailed overview of the image we've shared, which again, I just sent it the image. I didn't say anything and it autofilled this refer to the following content. This kind of visual storytelling is often used in stock photography, relationship articles, or AI generated scenes to convey themes like one partner being unaware of the other struggles, happy on the outside, hurting on the inside, relationship tension, or unspoken issues. The next thing I'm going to do is say, give me their names. And if we get like Elena or Leo or something, then there's really some like weird things going on inside of every AI, no matter who the lab is, because I've seen those names for this photo multiple times. Oh, Mike Thompson and Sarah. Never mind, for fun, try Derek Harland for the man and Melissa Harland. Okay. The man is Matt Dunovan.
The woman is Jenna Dunovan. They're a married couple in their early to mid-40s. Matt works in construction management and the eternal optimist.
Jenna is a high school guidance counselor who's been carrying a lot of stress lately, which explains the worried expression. And now we're going to turn this into one final front-end design task as I'm pasting in now the same prompt that I used in the Opus 47 test with the specific image saying, "This is the book cover for an upcoming book called Renovation or Resurrection.
You must design an insane website for the launch of this book with the cover displayed as well as a verbal narration of snippets from the book. interactive experiences make this a 10 out of 10 NYC publisher style knockout design that is high-tech and aesthetic. All right, so we did finally get it working. I just had to manually kind of put the photo in the same directory as the website file, which is acceptable, but I mean it should be able to figure out a way to get that image in there. Claude did base 64 and embedded it, I believe, so it is possible. But regardless, we have some nice moving particle effects in the background with the title prominently displayed right here. We can pre-order it for $28. And we have a nice tilt effect on the book cover right here, which does look good. Author name in the top right. Renovation or resurrection.
I've been seeing this tilt effect implemented more across a wide variety of different models from different providers. So, that's interesting. I do like that. 384 pages. A devastating luminous portrait of modern marriage.
Voss has written the book of the decade by the New Yorker. In one frozen moment, a husband smiling, a wife haunted, the entire architecture of a marriage is revealed. Renovation or resurrection asks the only question that matters.
When everything you built together is cracking, do you repair it or let it burn? Immersive storytelling. I totally forgot about this. We have to put in like the Oh, that's cool. Emotional X-ray. Click the cover to reveal the hidden truths. That's pretty good. We told it to put in some immersiveness.
Matt 94% joy. Jenna 87% pain. And when you click it, it has an effect.
Wait, wait, wait. What? Look at the bottom of the page when I click this book cover. Hidden truth revealed. Jenna has been planning her exit for 14 months. This is immersive AF. What would you do? Renovate or resurrect? All right, let's try. Let's see if this narration works.
>> The photograph captures everything.
Matt's arm around her shoulders, his smile wide and genuine.
>> And it's using accessibility the web GPU. in that single frozen second.
>> So that's very good. That did work.
Renovate. You chose renovation. You believe the marriage can be saved, but at what cost? Resurrect. You chose resurrection. And they both give us the nice confetti effect. Interesting. Early claim. What the world is saying. The book everyone will be talking about. 26 days till the official launch. This was nice. It was immersive and I'm pleased with that. So overall, that's going to conclude our first look and test of the Gro 4.3 beta. Overall, this is definitely a significant improvement over previous versions of Grock. 4.20 was a significant improvement over all of the preceding models, and this is the same case. I don't know how gigantic the leap is between 42 and 43 comparatively to how big the leap was from like 41 or 4 to 42. But regardless, we did get some pretty solid results here. And I would say I think overall the one that probably stands out the most is going to be this browser OS. The implementation of not only this Grand Theft Auto driving game, but the voice feature as the special feature where you can literally just click that and be like, "Open GTA drive." It's already open, so it won't do anything. But it also would be like, "What time is it?"
Oh, I turned the speaker off cuz the site wouldn't stop playing the book site. But that was all really impressive and we saw it earlier and the 3D game it placed in here. Beyond that, the other one that it gave us as well was pretty cool. So this was a very very promising browser OS a very good one as well with a more difficult test just because it has to make the 3D games. I would say of course how could we forget our skate game result which it first generated that asy game which was just perplexing and then we yelled at it a bit. It did apologize profusely. It was genuinely seemingly kind of regretful and felt bad. And then we got this which was it worked. It wasn't like a total wow fest.
I don't know why I put it that way, but it did give us something that did function. So, and we could actually hit NPCs and bail out. So, that was good to see as well. The subway station FPS was functional, but it was odd because of the way the camera logic was. So, there should have been a fixed Z height for the player and the recoil effect shouldn't have been directly on the camera or if it was it. No, it just should have been on the weapon, not on the camera. But, it did properly first create this map and then turn it into a little zombie game. So, that was cool to see. I mean, definitely improvements in capabilities all around. 3D Printer Sim had a funky UI. I kind of liked it. some front-end capability did show itself here where even in the Stevie Lapis portfolio website, I did like some of the effects on these cards and some of the other things here. I mean, just the way that the scrolling works, we see, look at the way this scrolls down till it stops and then it picks back up again when you're up, if you know what I mean.
This was not bad. All of this was cleanly arranged and fairly solid. I really personally very much liked the Steve the PC repair game. This was cool.
I love the cube particle like blow up effects when we actually destroy a computer or one of the enemies. Jumping worked. This ran well from like a frame rate perspective. And also it was just like decent to play around in. And it was fun which I liked. I think what else was there? We had the drum kit sim which was perplexing. Um it worked. It was just kind of aggro in terms of the autoplay tracks with specific emphasis on the heavy metal one. It was very very aggro. And that's basically it. So overall, I noticed a few things about this model in terms of the way that it decided to behave. One of them was definitely that this thinks for a lot longer and generates scripts that are much longer than what I've seen in testing previous Groc models. So some of these results really took their time, and I wasn't sure what to expect with that. As historically I've noticed that when doing like four model head-to-head comparison tests between like this Gemini Claude and GPT Gro is generally the quickest to finish producing its result. I don't know that that would be the case now. So it seems to take longer to think. Additionally to that, these scripts I noticed were pretty long some of them. So it was not afraid to really go all out in terms of script length and things of the sort. So definitely an improvement and it will be interesting to see what folks think of this when it becomes more accessible in terms of not being locked behind the $300 subscription, which I have heard, again, don't quote me on this, that it will kind of become more widely available to lower tier. So I'll be looking forward to that. And really, that's going to wrap it up. So if you have any questions, please feel free to leave them in the comments. And thanks for watching.
Related Videos
VALORANT's Latest 'Exclusive' Tier Bundle is Rough...
KangaValorant
17K views•2026-05-28
Flight Attendant Mocks Poor Looking Black Woman — Mid Air Announcement Exposes Her Real Power
SkyboundStories-b4r
184 views•2026-05-28
I FIXED My Friend’s Blown Turbo RX-8… Then Sold It
Cameron-RX8
134 views•2026-05-28
NewsWatch 12 at 5: Top Stories
NewsWatch12
1K views•2026-05-28
Simon Jordan & Danny Murphy deliver PREDICTIONS for Arsenal's Champions League FINAL with PSG
talkSPORTArsenal
6K views•2026-05-28
Botting is OUT OF CONTROL in Classic WoW (Again)...
SolheimGaming
108 views•2026-05-28
The "AI Job Apocalypse" is CANCELLED!
WesRoth
9K views•2026-05-28
STREET FIGHTER 6 - INGRID Story Walkthrough @ 4K 60ᶠᵖˢ ✔
RajmanGamingHD
12K views•2026-05-28











