Deepseek crushed everyone

0:00 Deepseek's latest model flew under the radar, and it is the best open- source model on the planet. The first thing I wanted to test combines a few different capabilities, including its ability to

0:10 generate 3D models, spatial awareness, internal knowledge the model has in its weights, and I think this is such a cool test. So, I said, build a 3D mechanical watch diagram so we can see exactly how

0:23 a mechanical watch works. Make it hyperrealistic and beautiful. when the user hovers over a piece, add information about what it's called and what it is used for. And here's what it

0:32 came up with. This is the Mechanica. And look how beautiful it is. And if you look closely, all of the pieces work. They fit correctly. These are tiny little pieces. And it looks 3D. You can

0:46 see the lighting. You can see the shadows moving as I'm moving the watch face. You can see the seconds hand, the minute's hand, the hour's hand. Everything looks really good. Now, if I

0:55 hover over any of these things, it tells me what it is. I can click over the hair spring. So, if I just hover there, it says hair spring, a blued spiral spring that returns the balance and sets the

1:06 watch's rhythm. If I click into it, it kind of highlights and centers it. So, the hair spring is a flat spiral of hardened steel, no thicker than a human hair. It stores the balances kinetic

1:18 energy at the end of each swing and gives it back on the return. So, I can say examine closely. It zooms into it. Look how awesome this is. It gives me some data about it. I can actually click

1:31 into each of the systems right here and then the overall parts of the watch right here. Go back to the overview and we can spin it around. We can look at the back. The one thing I don't see here

1:44 is the little weight on the back where you move your wrist and it moves around and charges the watch. I actually am forgetting in this moment what that's called, but nothing is here in the back.

1:54 So, that was a little disappointing, but otherwise, I think this is incredible. Now, check this out. If I click explode, it actually explodes it. And you can see each layer of the watch completely

2:06 exploded out and you can still hover over it. And of course, all of this was built in a single web page. And I just think this is absolutely beautiful. the detail, the spatial awareness, it's 3D

2:20 generation, the lighting, the shadows, everything looks really good with the exception of the back, which is the only improvement I would have liked to see. All right, the next thing I had it

2:28 create is a very popular multiplayer game called Fall Guys. And this is one that I've done a number of times before. So, while I show you this, I'm going to show you the other versions that other

2:39 models have created. And so, the prompt for this was quite simple. Build a Fall Guys clone. All right. So, this is called Bean Blitz and let's play. So, immediately I see it looks really good.

2:51 The 3D generation is really good. This is again just built with HTML. Oh, I got knocked off. Okay, so yeah, everything 3D looks really good. You can see some clipping of the different

3:06 objects. Uh, this doesn't actually do anything. So, it looks like it would be spinning, but it's not. So, that's okay. I can't fall in the acid right here. So, let's keep jumping.

3:19 Now, the hammers look good. Let's see if it actually knocks me off. It does. Boink. I can dive hitting the space bar. Okay, fell off. Now, we have these balls and I lost. So,

3:31 I think overall I'm really impressed with its ability to clone this game. It isn't as good as Opus 5.5, but it's quite close. And keep in mind, this model is a fraction of the price and

3:47 very, very fast. All right. Next, I wanted it to create a 3D model of the iPhone Duo, which requires it to go out on the web, understand the spec, understand the design, and actually be

3:58 able to recreate it in 3D space. And I think it absolutely crushed this test. So, here it is. And I think this is flawless. The Apple logo looks perfect. Look at that. You can even see the

4:12 reflection of the light in the screen. This side looks matted. It is a matte finish on the inside screen. All the buttons are in the right place. The hinge looks correct. Everything is

4:25 flawless. And by the way, I did this exact test with the brand new Kimmy model. Let me put that on screen because it shows Deep Seek is much better. So then you can actually fold it.

4:39 Now, there's a little issue with the folding animation because when you click it, it kind of goes off screen. So, that is an issue with this simulation, but let's fold it. There we go. And so, the

4:52 actual screen saver disappears for a second while you fold it. But overall, I think this looks incredible. And if you're thinking about using this DeepSk model in production for an app you're

5:04 building, you should definitely make sure it is production ready using the sponsor of today's video, Coreweave. Coreweave Arena lets you load test your AI before you ship to production. Here's

5:14 what that looks like in Forge. Coreweave's AI development platform. It starts with one request at a time, then five, then 10. I can see the response time, the throughput, and estimated

5:25 cost. When 10 requests came in, it took longer than five. But why? That's when I go to Arya, the AI assistant in Forge, to help me figure it out. It recommends reducing unnecessary processing for each

5:38 answer. Now, I'm going to test that before I launch. Arena is a guided program for testing Forge infrastructure or both with Coree's engineers. Coreweave is the essential cloud for AI.

5:50 Coreweave Arena gives you the insights you need to know on what will work, what will break, and what it takes to get your AI into production. So, apply for Arena Pass today. I'm going to drop a

6:00 link down below. Let them know I sent you and check it out. And this new DeepSseek model V4.1 Flash is also multimodal, which means it does accept images. And so, I uploaded this image

6:13 and it is a bunch of different pastries. And I wanted to see how well it understands what's in the image. And this is a really hard task. So, check this out. How many pastries are on the

6:24 counter? Count them individually. I went through the counter front to back. It took four minutes and 25 seconds in total. And counted every distinct item I could. Here's my tally. 137 individual

6:37 pastries visible on the counter. Number one, orange macarons. And it is in the pyramid and it counted 24. So if we look at it, this is what it's seeing right here. 1 2 3 4 5 6 7 8 9 10 11 12 13 14

6:52 15 16 22 23 24 25. Now, I count 25 visible, but 24 is very close, and it's arguable. Now, there are a bunch of macaroons that you cannot see in this image. And wouldn't it be nice if this

7:05 model could understand what is not being shown in the image just as well as what is being shown? I'm going to get back to that in a second. Here we go. Butter croissants, eight. Cream filled horned

7:18 croissants, nine. So, not only was it able to count every single pastry in the image, but it was actually able to name them all, which I find very impressive. Now, without me asking it, it said the

7:31 macarons, croissants, and brownies are stacked, so a handful are partly or fully hidden. Those three groups are the fuzzy ones. A defensible range for the whole counter is roughly 130 to 145. So

7:43 it immediately recognized that there are likely pastries behind the ones that are stacked that cannot be seen. So I asked it now estimate how many are on the table in total. Include the ones that

7:56 you might not be able to see directly. For example, there are probably more macaroons behind the ones that are visible directly in the pyramid. So here we go. Now it has the list of visible

8:07 and hidden. So it estimated another 21 hidden of those macarons. Here's butter croissants on the back stand. Let's take a look. So, here we go. 1 2 3 4 5 6 7 maybe seven and a half visible. And it's

8:21 saying there are five more that you cannot see that are stacked behind it. Look at its spatial awareness here. Macaroons are the big ones. It reads as a cone-shaped heap about 5 to six wide

8:33 at the base and 5 to six layers tall. A triangular front face of 24 implies a base layer of roughly 19 to 25 tapering up. So 45 total is much more likely than 24. Here we go. The croissant horn piles

8:49 2 to three layers deep. So I added about 60% to the visible count. I am beyond impressed with this. And the cool thing is the original count only took 4 1/2 minutes. And the follow-up count, it had

9:00 already done the calculation. It only took another 23 seconds. And by the way, this is on max reasoning. So stellar visual understanding with this DeepSeek model. All right, the next thing I want

9:12 to show you is how fast this model is. It has the moniker flash, which means it's going to be very fast, but it is surprisingly fast. So, what we're going to do, we're going to select DeepSeek

9:24 V4.1 and I'm going to turn off reasoning effort and I'm going to say write a 1,000word story in this chat window. Do not use tools. Now, the reason that I do that is

9:40 because even if you have the thinking effort set to off and you tell it to write a thousandword story, it's still going to think, okay, maybe I should spin up a text document. Maybe I should

9:52 do something else. All I wanted to do is output the text directly in this window. So, let's send it and let's see how fast it goes. Look at that. That's probably

10:04 125, 150 tokens per second. And that's not the fastest you can get it. In fact, I'm going to show you this in a little bit. This is artificial analysis. You can get this model up to 1,286

10:17 tokens per second from Lithos AI. Base 10 is serving it at 845 tokens per second. So, you can get very fast inference with this model. That is why I'm thinking this might be the current

10:30 best overall model on the planet. Not the highest quality, but close. But with its cost, its quality, and speed, I think it's number one. Okay. Next, I wanted to test its ability to A use a

10:44 browser, b understand how to specifically use YouTube, c look at a video and understand what a clippable moment in an interview is, and then d give me the correct timestamps. So, I

10:58 gave it my Sam Alman interview. I said, "Review the YouTube video and extract the top five most clippable moments. Make sure to give me exact timestamps and why you think it is worth clipping."

11:10 Then I just linked the video. So, it worked for 6 minutes and 24 seconds. And you know what's really cool with an open source open weights model? You can see all of the chain of thoughts, all of the

11:19 tool calls. So, the first clip it decided on, I haven't looked at a line of code in a year, and it gave me this specific timestamp, which is correct. I verified it. It gave me a link. It told

11:29 me the exact quote. I realize I have not read or even looked at a line of code in probably a year. I don't think I'll ever do that again. Or if I do, it'll be like a Revolutionary War reenactment when I'm

11:40 old and bored and trying to remember the glory days. And that's Sam Alman's quote. Why did it choose that? The CEO of OpenAI admitting he hasn't read code in a year is inherently newsworthy. And

11:52 the Revolutionary War reenactment line is a vivid, funny image that travels well outside the AI bubble. self-deprecating, zero setup required, perfect shorts, Tik Tok length, hook

12:03 text. Sam Alman hasn't read a line of code in a year. So, that is a very good choice in my opinion. Next, coders have never been busier. Another great choice. Three, the consciousness debate is a

12:17 backdoor to taking power from people. Another really good clip, one of my favorite from that interview. I get very nervous when people start using the AI consciousness debate to push into, well,

12:28 maybe people shouldn't always be in control of the future. The future has to be about people. And so, yeah, these are all clips that I would have personally chosen. And so, the next thing I would

12:37 probably give it, although I haven't done it yet, is to actually download the video and clip itself. And I think it can do that, but that'll have to be for another video. All right, here's the

12:48 next one. A full Microsoft Excel clone. And I want to say something about this before I share it with you. I loaded the DeepSk model locally. Now granted, I have a DJX station, which most people

13:01 don't have. It's very expensive, but you can get a quantized version of DeepSeek on a decently high-end consumer-grade GPU, and it makes me so excited. So, I ran this and built this completely

13:15 locally. So, all I said to do was build an Excel clone for the browser. It worked for 50 minutes and it built GridLine, an offline first Excel clone. So if I click into it, here it is. It

13:29 doesn't have the full functionality of Excel, but it is very good nonetheless. So if I double click into here, 23 + 44 = 67. I can say 44 here equals B7 + C7. And it does it correctly. It has bold. I

13:49 can drag this down. After dragging it down, it basically continued that formula into each of the cells that I dragged it to. So here I insert a chart and it has a beautiful chart. Very

14:01 simple. I can visualize another chart like so. There's another one. Drag and drop. Very easy. And next, I gave it the Rubik's cube test. Build a 3D simulation of a Rubik's cube. Scramble and solve.

14:14 Very simple prompt. And there's a reason I gave it a very simple prompt. Is it able to build the thing that I have in my mind? The thing that we all know what it should look like, but without me

14:24 having to explicitly tell it every single nuance of the task. And here it is. This Rubik's cube so far looks really good. All the sides look good. There's white under here, blue, orange,

14:39 yellow, green, red. Scramble. All the colors seem to be sticking to the appropriate sides. Everything still looks good after scrambling. We can actually see the moves scrambled right

14:48 here. And then if I click solve, everything looks good. The rotations look accurate. The physics of the cube looks accurate. And yes, it solved it. So, so impressive. This is an open-

15:03 source open weights model that you can download and do whatever you want with. And it is really good. Next, you should be building your own software at this point. I have built multiple apps

15:13 recently that I just needed in my life and it either didn't exist or the app that existed was garbage and I wanted to create something that was custom to me. One example is a PDF reader for Windows.

15:28 No, I do not want to use Adobe Acrobat. It is bloat. It is spammy and I do not need 90% of the features it has. So, I built my own and you can too. I had originally built this with Opus 5.5 and

15:42 I asked Opus 5.5 to give me a detailed prompt very different from what I'm used to doing with these tests but give me a detailed prompt and then I'll copy paste the prompt into DeepSeek. So I have it

15:54 deepseek high thinking settings build a complete PDF reader and annotation app called Folio from scratch. It's going to be using Electron React and TypeScript PDF.js for the PDF rendering. It told it

16:07 exactly how I want it to look and feel. Warm off-white light theme and near black dark theme, subtle borders, compact controls. Here it told it exactly how I want the PDF reading to

16:18 be. Drag and drop. You can highlight text, etc. And so after all that, it is a very detailed prompt. And here it is. Here's Folio. This is a beautiful simple PDF reader and editor. And it just

16:32 works. So some of the features, you have the pages, the table of contents. Here you can click around. You can hide that window if you want. You can zoom. All very like basic stuff so far. You can

16:44 draw on it. You can have different colors. You can move the drawing around. You can actually change the color just like so. You can add a text note. You have signatures. So, you can actually

16:54 draw a signature. You can upload a signature. You can take a picture. So you can like actually draw a signature on a piece of paper, hold it up to the camera, it will take a picture, flip it,

17:06 reverse it, and then you can use that signature anywhere you want. It also has dark mode, which I appreciate. You can just make any PDF into dark mode, so it's easier on your eyes. Saving, you

17:18 can export, fit within the page, you can rotate it clockwise, you can print it, everything you would expect from a PDF reader. And it was basically free to make. I mean, I ran it completely

17:30 locally. If you're paying an inference provider for this, you're probably paying less than a couple dollars to make a PDF reader that you can have for the rest of your life. It is so

17:43 impressive. So, from these demos, hopefully you can see why I'm calling this potentially the best overall model on the planet right now. There hasn't been another open- source model that I

17:53 have been this impressed with. And it's not just that it's fast and cheap and you can host it locally. You can get inference at over a,000 tokens per second. It's the quality. It is actually

18:06 high quality, actually high intelligence. And in fact, let me show you some of those benchmarks and stats. Now, here is one of the most important benchmarks you need to look at. This is

18:16 Deep Suite V1.1. And what you're seeing is this is the Paro Frontier on the quality cost combined metric. It is the best model. And so you can see that right here.

18:30 Here's a bunch of other models. Here's Opus 5, Fable 5, GPT 6.1 Soul, Astra 6 Soul, Gemini 3.8 Flash, GLM 5.3 Max, and up here in the top left, this is where you want to be. Higher up means better.

18:44 To the left means cheaper. and it is beating every single other model out there. I mean, this is super impressive. This model is the best bang for your buck. And you're not really sacrificing

18:57 all that much quality unless you're doing the most complex tasks ever. Here are some other benchmarks. We have Humanity's Last Exam. It's coming in at a 42, which is right on par with Kimmy

19:08 K3 and GLM 5.3. Here's Terminal Bench coming in at 42. Automation Bench Agents last exam. It just scores really well. But of course, benchmarks only tell part of the story. And now that you've seen

19:20 it actually in action, you should be using this model. And I want to show you one of the most important metrics that you need to be looking at anytime you see a new model, which is cost per

19:30 intelligence task. And that combines the cost of the model with how many tokens it takes to complete a given task. And so you want the lowest cost and you also want the highest density of intelligence

19:43 which results in less tokens being used per task. And so at the very top maybe unsurprisingly Claude Fable 5.1 Max at $763. I'll show you a couple others. Quen 3.8

19:56 Max at $5. Very expensive. Here's Astra at 3.26. Here's Kimmy K 3.1 Pro at $211. This model just came out and I think Deepseek is better than it. And all the way down

20:11 here, look at this. 26 cents per task completed. That is so inexpensive and it's a very good model. And speaking of price, here's what you can get the model at on Lithos AI. You can get it

20:26 for 60 cents per million output tokens, 15 cents per million input, and it's free if it's a cash hit. Free. Here's base 10 probably with the overall best offering right now because remember it

20:39 gets about 800 tokens per second and it's a$120 per million output, 30 cents per million input, one penny per million cash hit tokens, one penny. I'm so impressed with Deepseek. Keep shipping

20:53 the goodness, Deepseek. Thank you. And I mentioned Kimmy K3.1 a few times and it's the newest model from Moonshot and it scored higher overall in the benchmarks. But I personally think

21:05 DeepSk is much better. You should make the comparison yourself. Go watch my Kimmy video