start	end	text
0	5740	So DeepSeek version 4 Pro is officially out today. Now you might be confused because this model was
5740	10680	technically out but it was in general availability but this is the official release meaning that they
10680	15200	have trained the model a little bit more and produced a stronger version of the model that
15200	19900	is available today. Now if I were to summarize today's release in one simple sentence it is that
19900	25000	the price to performance is becoming a very very important thing because not only do we have a new
25000	30660	model from the DeepSeek team, the SpaceX team also dropped Grok 4.6 and both of these models
30660	36520	are competing with the best frontier labs at a fraction of their cost. Let's start with Grok 4.6
36520	41180	and one thing I'm going to say is that I'm genuinely surprised by the SpaceX team. I'm not
41180	45860	trying to glaze them or Elon Musk or anything. I was just very critical of this lab. In 2025,
46240	51380	they dropped Grok 4 and other models like that but I wasn't really you know mind blown with their
51380	56240	performance because they're pretty they're pretty subpar compared to any of the other labs out there
56240	61680	but in 2026 it looks like things have kind of changed a little because grok 4.5 was quite
61680	68280	competitive based off what it cost and grok 4.6 is actually not too bad if we take a look at the
68280	73300	benchmarks for example if we start with the artificial analysis intelligence index this
73300	80080	model achieves a 61 and fable 5 is at 62 now what's really important to remember is that once again
80080	85460	this model is quite cheap compared to Fable 5. This is about $2 per million input tokens
85460	91900	and $6 per million output tokens. And Fable 5 sits at $10 per million input tokens and $50
91900	96700	per million output tokens. And this is why you start to appreciate this release a little bit
96700	102240	more. It might not beat the performance of the best models. It's matching them. Even GPT 5.6
102240	107540	sold, which is a pretty capable model, and this is set at max on the artificial analysis index.
107540	114660	this model achieves 61 and grok 4.6 61 so it ties it and it's much cheaper and then even on all of
114660	120420	these other benchmarks for example the gdp valve one it actually beats fable 5 which is at 1741
120420	126900	and then gpt 5.6 it's at 1728 i'm not sure why they didn't choose opus 5 as well but i guess
126900	131460	they wanted to choose the quote-unquote strongest model lineup from each lab and they chose fable
131460	136140	5 for anthropic which is fair and then deep software engineering one which is a critical
136140	143720	benchmark this model doesn't beat fable 5 or gpt 5.6 soul but it gets close to it it's 65.9 and
143720	149040	fable 5 sits at 70 but if you're getting results that are pretty close and the model is five times
149040	154080	cheaper i wouldn't be too disappointed with this result and then same thing with cursor bench 3.2
154080	162240	the model achieves 69.9 funny number and fable 5 sits at 70.5 so once again closer and grok 4.6
162240	168200	beats GPT 5.6 soul. The same thing with the Frontier Code, it gets close to Fable 5 beats
168200	173940	GPT 5.6 soul. So yeah, this model is actually available in Cursor. So the partnership with
173940	178960	Cursor or I guess the acquisition has really helped SpaceX make some strides in the AI space
178960	184120	this year. And Grok build is something that, you know, maybe not a lot of us have been using so
184120	189480	far, but it's probably going to be another platform like Codex or Cloud Code that we start to use.
189480	194580	but obviously cursor is quite strong as well so you have options available for you to use them in
194580	200240	both and one thing to note is that they're offering two times usage inside grok build and cursor for
200240	205320	the first week so if you just want to try it out see what you feel about it then you know it might
205320	209760	be worth trying it out right now because you get double the usage in the first week now if we were
209760	214520	to take a look at some of the outputs that people have been generating with grok 4.6 what we're
214520	220180	looking at right now is a Falcon 9 booster return sequence simulation and this was done in a single
220180	227620	HTML file and as I said if you expected to get this type of output from Grok in 2025 you would
227620	232140	be kind of surprised because you wouldn't expect something like this to be generated with Grok but
232140	235960	now it looks like we have to start taking the Grok team a little bit more serious because
235960	241160	this output is quite competitive and obviously we're just looking at a simulation and we're just
241160	245940	basing it off of a visual representation but if the model is able to produce something like this
245940	251340	consistently then I would expect a lot of people to start adopting Grok because number one it is
251340	255660	cheaper than the other labs at least at the moment because we don't know if this pricing strategy is
255660	260360	going to be sustainable for the SpaceX team in the long run but at least for now their models are
260360	265280	definitely cheaper compared to the others and as I mentioned this model excelled at the artificial
265280	272580	analysis index. This model jumped to probably number four model. It's tied to GPT 5.6. It's
272580	276560	pretty much similar. So you could say number three as well. But the models before that are
276560	283540	Fable 5 and Opus 5, which are only above the model by about a 1% or a 2% difference. So yeah,
283580	288640	even on this intelligence index, which if you're not familiar with, has nine evaluations. So on
288640	294220	all of these evaluations, it's kind of matching almost Fable 5 performance, which is crazy to see.
294220	299760	Before we continue, we just launched the Universe of AI newsletter. If you want to stay on top of
299760	305060	AI news without having to hunt for it, link is in the description. Don't miss out. And what you see
305060	310980	on screen right now is a racing game that Grok 4.6 built. And based off of this post, the model took
310980	315780	about one minute and it was a five word prompt, which was a create a simple racing game in HTML.
316280	321200	So if you're able to generate something like this easily using Grok 4.6, I think a lot of people
321200	327160	will be happy. And this is a more detailed analysis of what it costs to run GPT 5.6 on
327160	332120	the artificial analysis index and what it produced, meaning the output. Both of these models, if you
332120	338120	remember, scored 61 on the artificial analysis index. Now to run the whole test with GPT 5.6,
338480	345500	it costs about 2.8k. And then with Grok 4.6, it costs about 1.1k-ish. And this tells you that
345500	350420	you're getting similar level of performance at half the cost. So yeah, this is a big release for
350420	354280	the SpaceX team because they just proved once again that they are a lab that you seriously
354280	359900	start into considering especially in 2026 and I'm going to talk more about DeepSeek version for Pro
359900	365600	GA but basically what we're seeing today is that both of these releases kind of emphasize the fact
365600	370520	that performance and all above that is the price at what you're getting for that performance is
370520	375880	becoming more and more important for all users because we see many labs now focusing on creating
375880	381380	the best model at the cheapest cost. Last year in 2025, most of the labs were just focused on,
381440	386680	I would say, creating the strongest model. Yes, cost was important, but I think most of the times
386680	391620	the frontier labs, meaning OpenAI Anthropic, were kind of more lenient on that fact because they
391620	398020	didn't have as strong of a competition. Intelligent models that are maybe not always ahead of OpenAI
398020	402560	Anthropic, but match their performance at a fraction of the cost. So yes, price to performance
402560	409800	ratio is becoming a critical I would say indicator in 2026. Now this is the updated benchmark chart
409800	414120	after the release of the new model and one thing you'll see across the board is that it matches
414120	418360	the top level performance of many of the models. The one thing interesting over here is that they
418360	422960	haven't put Opus 5 here for some reason. There is Fable 5 here that we can compare this model
422960	428900	against but one thing you'll notice is that DeepSeek, remember this model costs 43.5 cents
428900	434940	per million input tokens and 87 cents per million output tokens while the other models all over here
434940	440140	are way more expensive than that. So the first thing if you look at Terminal Bench 2.1 the model
440140	446126	scores 87.9. The older version of the model was 72.1 and the flash version was
446126	449526	of the model was 72.1, and the Flash version, which we got last week, was 82.7. And what's
449526	457146	crazy is that Fable 5 is 88. Yes, 88. So this model is only 0.1% behind Fable 5. And then on
457146	463606	the Cyber Gym which is Cybersecurity the model actually beats Fable 5. Fable 5 sits at 83.1
463606	469546	while DeepSeek version 4 Pro the new one that we got today is at 83.3. Now if this tells you
469546	474706	something is that this model is once again really geared at Cybersecurity. We saw the Flash model
474706	478966	also be geared towards Cybersecurity and becoming a model that was quite capable in that area
478966	484986	and we're seeing the same thing today with the new version 4 Pro which sits at 83.3 outperforming
484986	490486	Fable 5 on the benchmark. On deep software engineering this model is not beating Fable 5
490486	497786	because Fable 5 sits at 70 but the model achieves 62.7 which is ahead of GLM 5.2 and a bit behind
497786	504186	Kimi K3 which sits at 67.5 but one thing again this is a big jump compared to the preview version
504186	509686	of the model which was at 12.8 so yeah they have really trained this model and have improved it on
509686	514706	the back end because we can clearly see in this deep software engineering benchmark. And there's
514706	520266	couple of other benchmarks but one key benchmark that this model excels at is the automation bench
520266	528346	where the model achieves 31.8 and fable 5 29.1 so once again the model outperforms fable 5 now this
528346	533166	is a fraction of the cost as i mentioned this is 43 cents and 87 cents for input and output
533166	539606	respectively versus fable 5 which is at 10.50 so yes if we look at the terminal bench for example
539606	545106	we are seeing this model achieve a result which is 0.1 behind the best model out there fable 5
545106	552766	and the pricing is insane to look at 43 cents versus 10 dollars 87 cents versus 50 dollars per
552766	559446	million input tokens so when we turn that into a fraction this model is 57 times cheaper and earlier
559446	564586	last week i think i made a video about how deep seek version 4 pro was going to achieve a result
564586	568866	like this because there was an investor report that leaked and when i saw the pricing where it
568866	574206	said that it was going to match Fable 5 or even outperform it and be at 57 times cheaper per cost.
574366	578906	When I was reading their investor report, I didn't really believe it. But today it is proven because
578906	583786	at least on the benchmark so far, yes, I'm just saying the benchmarks, we are seeing similar level
583786	589086	performance. Now, if this model actually performs like that in production, it's a little too early
589086	593506	to tell yet. We still would have to give a couple of weeks and see how it's performing in the long
593506	598206	run. Because sometimes on the release date, the models perform quite good. But over time,
598206	602406	they kind of deteriorate in their quality. So we hope DeepSeek doesn't do that. Historically,
602486	606726	they haven't done that. But let's just see. Just going to put it out there because right now we're
606726	611886	basing this off of benchmarks. But that's it for today's video. Make sure you guys are subscribed
611886	617666	to the channel. Follow our new newsletter as well at universeofai.beehive.com. As well,
617726	622906	subscribe to the main channel, World of AI, and support us on X by following the Universe of AIZ
622906	625706	as well. Until then, I'll see you guys in the next video.