DeepSeek V4.1 Flash: The Best Open Source Model Tested
A hands-on run through 3D builds, image counting, clip-finding and a local Excel clone, ending on speed and price charts.
Posted
today
Duration
Format
Review
hype
Views
97K
1.9K likes
57 · 43
Big Idea
The argument in one line.
DeepSeek V4.1 Flash is close to frontier quality at a fraction of the cost and with very high speed, which makes it the best overall model for most tasks.
Who This Is For
Read if. Skip if.
READ IF YOU ARE…
A developer deciding which model to use for coding and app-building agents who wants real demo results instead of benchmark tables.
Someone paying frontier prices for every task who wants to know whether an open-weights model can replace most of that spend.
A builder curious about running a large model locally on a high-end GPU and what it can actually produce.
SKIP IF…
You need rigorous, reproducible evaluation data; these are single-run, hand-judged demos.
You only do the most complex reasoning tasks, where the host says the quality gap to top models still shows.
TL;DR
The full version, fast.
The host tests DeepSeek V4.1 Flash on a 3D mechanical watch, a Fall Guys clone, a foldable phone model, counting pastries from a photo, finding clippable moments in an interview, and a locally built Excel clone, Rubik's Cube and PDF reader. Most results are strong, with small flaws like a missing rotor on the watch back. He then shows speed numbers (up to 1,286 tokens per second from some providers) and cost-per-task charts putting it at about 26 cents per task against several dollars for top models. His conclusion is that it is the best overall model once quality, cost and speed are combined.
Free for members
Chat with this breakdown — free.
Sign in and you get 23 free chat messages on us — ask for the hook, quote a framework, find the exact transcript moment, generate a markdown action plan. Bring your own key when you want unlimited.
A hyper-realistic interactive watch with hover labels, exploded view and one missing part on the back.
02:19 – 03:46
02 · Fall Guys clone
A playable 3D obstacle game built in HTML, close to a top model but not quite.
03:46 – 04:58
03 · Foldable phone in 3D
A spec-accurate 3D foldable phone, with a minor fold animation issue.
04:58 – 06:13
04 · Sponsor: CoreWeave Arena
Load testing an AI app before production.
06:13 – 09:08
05 · Counting pastries from an image
Multimodal counting plus an estimate of hidden items behind stacks.
09:08 – 10:45
06 · Speed test
A 1,000 word story with reasoning off, then provider speed numbers.
10:45 – 12:43
07 · Finding clippable interview moments
The model browses YouTube and returns timestamped clips with reasons.
12:43 – 14:58
08 · Local Excel clone and Rubik's Cube
A spreadsheet built over 50 minutes locally, then a scramble and solve cube.
14:58 – 17:58
09 · Building your own software: Folio PDF reader
A detailed prompt from a stronger model produces a full PDF reader and annotator.
17:58 – 21:11
10 · Benchmarks, cost and pricing
Pareto chart, cost per task, and provider pricing.
Atomic Insights
Lines worth screenshotting.
Open-weights models let you read every chain of thought and tool call, which makes agent failures far easier to debug.
A strong test prompt is a simple one, because it shows whether the model can build the thing you picture without you spelling out every detail.
Cost per completed task, not price per token, is the metric that matters, since it also counts how many tokens a model burns.
Asked to estimate hidden pastries, the model reasoned about stacked geometry and added about 60 percent to the visible count.
A follow-up estimate that reused earlier work took 23 seconds, against 4 minutes 25 seconds for the first full count.
Having a stronger model write a detailed prompt, then pasting it into a cheaper model, is a practical way to cut build costs.
Prompt caching can drop input prices to a penny per million tokens, so cache-friendly workflows change the real cost sharply.
The same model can run at 125 tokens per second in one chat and over 1,000 on another provider, so host choice matters as much as model choice.
A quantized version of a large open model can run on a high-end consumer GPU, so local app building is now realistic.
Custom single-purpose apps, like a PDF reader without the bloat, are now cheap enough to build instead of tolerating bad software.
Takeaway
Judge a model by cost per task
WHAT TO LEARN
Real build tests, speed and cost per completed task tell you more about a model than a single benchmark score.
Simple prompts reveal whether a model fills in the details you picture, which complex prompts hide because they spell everything out.
Open-weights models expose every chain of thought and tool call, so you can see why an agent made each decision.
Multimodal models can estimate hidden items from visible geometry, but treat the numbers as ranges and check them by hand.
A follow-up question that reuses earlier work can finish in seconds, so ask for estimates after the first expensive pass.
Provider choice can change speed from about 125 to over 1,000 tokens per second for the same model.
A quantized open model on a high-end consumer GPU can build a full app locally for almost no marginal cost.
Have a stronger model write a detailed prompt once, then run it on a cheaper model to cut the build cost.
Cost per completed task combines price and token use, so it is a better comparison than price per million tokens.
Cache hits can be billed at a penny per million tokens, which changes the real cost of long agent sessions.
Custom apps built for one need often beat bloated commercial software, and building them is now cheap.
Glossary
Terms worth knowing.
Open weights
A model whose trained parameters are published so anyone can download, run and modify it, instead of only calling it through a company's API.
Pareto frontier
The set of options where you cannot improve quality without paying more, or cut cost without losing quality. Models on the edge are the best trade-offs.
Cache hit
A request whose input the provider has already processed and stored, so it is billed at a much lower rate than fresh input.
Quantized model
A version of a model with lower-precision numbers, which shrinks memory needs so it can run on smaller hardware at a small quality cost.
Tokens per second
How many chunks of text a model produces each second, the standard measure of generation speed.
Reasoning effort
A setting that controls how much the model thinks before answering. Higher values cost more time and tokens but can improve results on hard tasks.
Resources
Things they pointed at.
04:58productCoreWeave Arena
10:00toolArtificial Analysis
20:20productLithos AI
20:20productBaseten
Quotables
Lines you could clip.
00:00
“DeepSeek's latest model flew under the radar and it is the best open source model on the planet.”
“26 cents per task completed. That is so inexpensive.”
concrete number against a $7.63 top model→ newsletter pull-quote↗ Tweet quote
The Script
Word for word.
Read-along
Don't just watch it. Burn it in.
See every word as it's spoken — crank it to 2× and still catch all of it. The same dual-channel trick behind Amazon's Kindle + Audible.
17px
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 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 a mechanical watch works. Make it hyper -realistic 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 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 see the lighting. You can see the shadows moving as I'm moving the watch face. You can see the seconds hand, the minutes hand, the hours hand.
Everything looks really good. Now, if I hover over any of these things, it tells me what it is. I can click over the hairspring.
So if I just hover there, it says hairspring, a blued spiral spring that returns the balance and sets the watch's rhythm. If I click into it. It kind of highlights and centers it.
So the hairspring is a flat spiral of hardened steel, no thicker than a human hair. It stores the balances kinetic 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 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 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.
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 exploded out and you can still hover over it.
And of course, all of this was built in a single webpage. And I just think this is absolutely beautiful. The detail, the spatial awareness.
It's 3D 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 to 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 gonna show you the other versions that other 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. 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 objects. 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.
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 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 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 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 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 flawless.
And by the way, I did this exact test with the brand new Kimi model. Let me put that on screen because it shows Deep Seek is much better. Then you can actually fold it.
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 actual screensaver disappears for a second while you fold it. But overall. I think this looks incredible.
And if you're thinking about using this DeepSeq model in production for an app you're 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 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 cost.
When 10 requests came in, it took longer than five. But why? That's when I go to Aria, the AI assistant in Forge to help me figure it out.
It recommends reducing unnecessary processing for each answer. Now I'm going to test that before I launch. Arena is a guided program for testing Forge.
infrastructure, or both with CoreWeave's engineers. CoreWeave is the essential cloud for AI. 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 link down below. Let them know I sent you and check it out.
And this new DeepSeq model, v4 .1 Flash, is also multimodal, which means it does accept images. And so I uploaded this image. 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 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 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. One.
two, three, four, five, six, seven, eight, nine, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 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 model could understand what is not being shown in the image just as well as what is being shown. Now I'm gonna get back to that in a second. Here we go, butter croissants, eight.
cream filled horn 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 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 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 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 and hidden. So it estimated another 21 hidden of those macaroons.
butter croissants on the back stand. Let's take a look. So here we go.
One, two, three, four, five, six, seven, maybe seven and a half visible. And it's saying there are five more that you cannot see that are stacked behind it. Look at its spatial awareness here.
Macarons are the big ones. It reads as a cone shaped heap about five to six wide at the base and five 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, two 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 four and a half minutes and the follow -up count, it had 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 deep seek model. All right, the next thing I want 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 deep seek V 4 .1, and I'm going to turn off reasoning effort. And I'm going to say, write a 1000 word story.
in this chat window, do not use tools. Now, the reason that I do that is because even if you have the thinking effort set to off and you tell it to write a thousand word story, it's still gonna think, okay, maybe I should spin up a text document. Maybe I should do something else.
All I want it 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 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 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 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 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 gave it my Sam Altman 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, then I just linked to the video. So it worked for six 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 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 the specific timestamp, which is correct, I verified it. It gave me a link.
It told 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 old and bored and trying to remember the glory days. And that's Sam Altman's quote. Why did it choose that?
The CEO of OpenAI admitting he hasn't read code in a year is inherently newsworthy, and the Revolutionary War reenactment line is a vivid, funny image that travels well outside the AI bubble. Self -deprecating, zero setup required, perfect shorts TikTok length, hook text, Sam Altman 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 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, 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 probably give it, although I haven't done it yet, is to actually download the video and clip it.
itself. And I think it can do that, but that'll have to be for another video. All right, here's the next one, a full Microsoft Excel clone.
And I want to say something about this before I share it with you. I loaded the DeepSeek model locally. Now, granted, I have a DJX station, which most people 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 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 doesn't have the full functionality of Excel, but it is very good nonetheless. So if I double click into here, 23 plus 44 equals 67.
I can say 44 here equals B7 plus C7, and it does it correctly. It has bold. I 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 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.
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 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, 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 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 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 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. 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 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 DeepSeek. high thinking settings, build a complete PDF reader and annotation app called Folio from scratch. It's gonna be using Electron, React and TypeScript, pdf .js for the PDF rendering.
It told it 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 be, drag and drop, you can highlight text, et cetera. 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 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 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 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, 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 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 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 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 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 thousand tokens per second, it's the quality.
It is actually 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 DeepSwee v1 .1. And what you're seeing is this is the Pareto frontier. On the quality cost combined metric, it is the best model.
And so you can see that right here. Here's a bunch of other models. Here's Opus 5, Fable 5, GPT 6 .1 Sol, Astra, 6 Sol.
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, 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 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 K3 and GLM 5 .3. Here's Terminal Bench coming in at 42.
Automation Bench, Agent's Last Exam. It just scores really well. But of course, benchmarks only tell part of the story.
And now that you've seen 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 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, which results in less tokens being used per task. And so at the very top, maybe unsurprisingly, Claude Fable 5 .1 Max at $7 .63.
I'll show you a couple others. Quen 3 .8 Max at $5, very expensive. Here's Astra at 326.
Here's Kimi K3 .1 Pro at $2 .11. This model just came out and I think DeepSeek is better than it. And all the way down 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 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 gets about 800 tokens per second, and it's $1 .20 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 the goodness, DeepSeek. Thank you. And I mentioned Kimi 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 DeepSeek is much better.
You should make the comparison yourself. Go watch my Kimi video here.
The Hook
The bait, then the rug-pull.
A model that flew under the radar, the host says, is now the best open source model available. The rest of the video is a string of live builds meant to back that up.
Frameworks
Named ideas worth stealing.
15:28concept
Strong model writes the prompt, cheap model builds
Ask a top model for a detailed build prompt, then run it on a cheaper open model to get the same app at far lower cost.
Add Modern Creator as a preferred source and Google shows you more of our breakdowns in Search, Top Stories, and AI Overviews. It only changes what you see, and you can undo it in your Google settings anytime.
Add to Preferred SourcesOpens your Google source preferences with us pre-loaded. Tick the box and you're done.
A tour of TypeSafe AI's new decision-making model, Jev, and the eight things people already built with it in its first week: email sorting, ad blocking, instant video clipping, and more.
ChatGPT co-inventor Diogo Almeida ships a model that's 200x faster and effectively free, but Matthew Berman's own chess test shows it losing on skill, only surviving the clock.
The mystery model that took over OpenRouter turns out to be a Chinese open-weights release that nearly matches frontier intelligence for a few cents a task.
A 'dot' release plays out like a full generational leap: two five-to-seven-day unsupervised coding runs, a sponsor benchmark, and a live pricing and capability standoff against a rawer, higher-ceiling rival model.
A first-look review of Claude Fable 5 and Mythos 5 from someone with early access: benchmarks, pricing, firsthand quirks, and two live multi-agent demos.
A 45-minute walk through Anthropic's internal data showing AI crossed from coding assistant to primary engineer — and a frank read on what that means for humans.
00:31
00:49
01:11
01:27
01:43
01:59
02:15
02:27
03:02
03:18
03:34
03:48
04:26
04:36
05:14
05:25
05:47
05:50
06:15
06:45
07:01
07:14
07:32
07:48
08:04
08:15
08:36
08:52
09:08
09:24
09:39
10:11
10:27
10:43
10:59
11:31
11:46
12:02
12:18
12:34
12:50
13:06
13:26
13:54
14:07
14:25
14:41
14:57
15:08
15:29
15:45
16:17
16:26
16:42
17:09
17:20
17:43
17:52
18:11
18:24
18:40
19:16
19:23
19:43
19:59
20:20
20:31
20:50
21:03
10:38
12:31
18:57
08:54
28:16
44:52