NEW Claude Sonnet 3.7 Beats ALL Models!🤖 OpenAI, Grok 3, Gemini 2.0 Pro - Best Coding AI Agent
In this video
- 00:00Introduction to Claude Sonnet 3.7 Release
- 00:32Key Features: Hybrid Reasoning Mode & Claude Code
- 02:00Pricing & Benchmark Performance of Sonnet 3.7
- 04:50Generating a React Landing Page with Sonnet 3.7
- 06:42Building a CRM Application with Extended Thinking Mode
- 08:29Reviewing the Generated CRM Dashboard & Features
- 09:25One-Prompt AI Code Generation Examples
Links from the video
Transcript
2,727 wordsAll right guys so on this channel we've covered many different AI coding tools as well as different new llm releases and it's been really heating up since 2025 and especially these last few weeks it seems like every other week every other day there's new llms there's new tools and it's really just getting crazy and things are starting to accelerate so as many of you guys probably already know Claude Sonet 3.7 just got released today and for a while clad Sonet 3.5
Even though there's been grock 3 there's been deep seek gar y there's been 03 mini 01 us developers were mainly still using Cloud Sonic 3.5 for many different development tasks just cuz it is so good and they put out claudet 3.5 a long time ago but now with Cloud Sonic 3.7 I can definitely say the difference is very noticeable all right guys so all links I cover in today's video will be linked down below here I'm on the announcement page of
What anthropic put out for claw Sonet 3.7 and Claude code so not only this Claud Sonic 3.7 is their most intelligent model to date and the first hybrid reasoning model on the market so you can use it in just normal mode or in reasoning mode now Sonic can produce near instant responses or extended step-by-step thinking that is made visible to the user so this is awesome because you know even in open AIS uh it look looks a little bit different from
What the actual thought process is and they're actually hiding it but here um from what I can see it looks like they're at least showing a lot more with the actual thinking process so this is nice to see as well something like the Deep seek R1 effect something like deep seek R1 I don't know if it's exactly or if it's abstracted a bit and the cool thing is API users also have fine grain control of how long the model can think
For now if you're a front-end developer this is definitely the llm you want to be using and like I mentioned they release CLA code here which is essentially like a AI coding terminal tool something similar to ader and what's even crazier is that claudon 3.7 is available on all clawed plans including the free plan the pro and the team and Enterprise now you're of course going to get more rate limits on the free plan and whatnot but it's still awesome that
Anyone could access this one downside is that it does have the same API price as its predecessor 3.5 which is $3 per million input tokens and $15 per million output tokens which includes thinking tokens so so it is pretty expensive comparative to other models but of course you know you get what you pay for this is the best coding model in the world right now and the benchmarks on this thing are very very impressive to say the least with we can
See claw 3.5 here on software engineering swe bench verified we can see that all the other models before it like 3.5 01 03 High uh deep seek R1 are all very similar around that 49% and then boom Sonet 3.7 comes in here getting a whopping 62.3% and then this little filled in thing I don't know what this custom variable that they added here is but pretty impressive and then also this is a very important Ben bench mark as well the agentic
Tool use T bench you can see here that Claude Sonet 3.7 is taking the lead on this as well and then here we have more benchmarks so 3.7 Sonet extended thinking and then 3 Sonet no thinking right here can see graduate level reasoning um you know pretty high at grock is doing really nice here with they're 80 and 84% so it's actually the highest and then plots on at 3.7 looks to be in third 03 mini high is actually above as
Well and I'm not going to go through all the benchmarks here but we can see you know it's doing if you go through pause it you'll see that it has some very impressive benchmarks agentic tool use the Cloud's always been really good at following instructions this is pretty cool too as well we can see that their you know road map here Claud assists so Claud helps individuals do their current work better making each person their best version of themselves in 2024
CL collapse at so Claude does hours of independent work for you on par with experts depending what every person on or team is capable of and that's in 20124 and 2025 and then Claude Pioneer so Claude finds breakthrough solutions to challenging problems that would have taken teams years to achieve in 2027 so you know if their road map is accurate that's only two years away and and I definitely don't doubt that this is where we could be in that time so
This is another interesting Benchmark right here this is claud's models playing Pokemon so you can see Milestone progression over time and you can see that CLA Sonet 3.7 is doing the best over models like CLA Sonet 3.5 and Claud son 3.5 new which people are calling CLA sonit 3.6 okay so it's very simple to go ahead and start using CLA son 3.7 you just go to claw. a and make sure that you're selected uh for CLA on 3.7 you'll have the
Option down here to either change to normal okay or extended which is their thinking model right here all right so here I'm saying please code me out a beautiful full react landing page for my AI automation business we're going to click on send and we can see that it has started coding right here very quick okay it's been going for a long time like it's writing a lot of code right here like very indepth holy moly okay so this is the
Website we can see right here so automate AI transform your business with AI automation watch demo start free trial okay we got our features right here industry Solutions okay all the different Industries what clients are saying transparent pricing right here okay we got our footer and everything all in all it coded out this landing page in first prompt one shot and it did a pretty good job I'm going to try the same prompt with Claud Sonic 3.5 and this is the
3.6 the one that was launched in October 2024 I have a feeling it's not going to be as long yeah so it was a lot quicker definitely nowhere near as long or as in depth as you can see this is what we got here so ready to automate your success little footer down here some stats watch CH R Solutions and doesn't even have a a nav bar right anything that is compared to this we can see we got our logo up
Here this one doesn't have any logo or anything much better um transform your business with AI automation it's funny that it got the exact same thing streamline your operations boost your productivity and unlock so it's using the exact same copy which is kind of funny but uh and it's similar colors but it's definitely a big upgrade so using CLA onet 3.7 within here you're going to definitely notice a big difference if you're used to to using CL on 3.5 with coding
All right so now I'm going to say create an in-depth CRM application with react it should have contacts page a built-in task manager calendar and project management system make it very beautiful and now we are using the extended thinking model and we're going to send this okay so we can see the thinking process right here if we click here can see the user is asking me to create an in-depth CRM application here's what it needs to have boom boom boom and
Then there are they specifically want it to be beautiful this is what I'm going to do here's the plan so it thought very quickly okay so it is going and it's generating a lot of code that's one thing I'll say about 3.7 is it seems like it can generate end to endend front-end apps very very easy or at least the first iteration and it will generate a lot more code than Sonet 3.5 did on the initial run okay this is very
Long like this is even longer than the first one it's probably cuz it has a lot of features but curious to see how this one's going to look okay it's still going like this is a very very long generation I've never actually seen a generation this long you can see here last edited 3 minutes it's been over 3 minutes so it's just been spinning this up okay this is honestly almost been 5 minutes now it's getting pretty crazy that it's generating
This long I I can't believe this but hopefully it works within the artifact okay so I just got Claude artifact failed to load I can see here Claude has hit the max length for a message and has paus its response you can right continue to keep the chat going I'm just going to say just finish the app Quick so it loads an artifact all right so we can see that it thought here and now it's making edits okay so it said
It did the task content now it's doing the settings right here to finish up apparently okay so it should be done artifact failed to load I'm going to say try fixing with claw okay so I actually had to tell it to just finish it up make it shorter and less in depth but still in depth because it was just generating so much code it was actually you know breaking the artifact cuz it would hit its actual limit here as you can
See we got a beautiful looking dashboard right here contacts you know add contact tasks right here new tasks calendar okay so you can see projects so really nice stuff with one shot prompts you can even toggle on Dark mode and light mode on even though it's not perfect but um this is really cool I think um for one prompt a lot of people are building some really cool stuff like here is a Twitter thread I'll link down below you can see
Here with this test it created a Minecraft clone in one prompt and made it insanely playable in artifact so this is pretty insane if you can see right here and then here this guy got a one prompt Sims game right here in artifacts single thing it did here in just a second but you can see it did in fact generate 26 files it technically created 23 and modified uh three ones this is literally just from a xjs you know create CLI
Type deal so I didn't write any of this code beforehand and one thing I want to do let me save this I wanted to show you guys all these new files so let me just save this to this Branch really quick what I'm going to do is I'm going to swap back to 3.5 just cuz I want to show you what the same prompt with Claude 3.5 Sonet did um this is all it did so it got some of the footer
Stuff um SAS name it kept like the basic nextjs page here um you can see it doesn't have any of these pages I don't even think it has the off Pages like it literally just did I want to say it did like five files I already committed this um but it literally was just like five files and so if we jump back to the three and one shot from literally the prim it literally yeah so it looks like everyone is kind
Of talking about this huge one shot you know capability it has where it can generate really good stuff with just one shot so so one thing I'll be doing a lot more videos around Claude Sonet 3.7 as well as how to use Claude code as well as using Claud Sonet 3.7 incursor if you haven't been using an incursor yet I would definitely recommend because honestly it makes it just feel like magic like I was just working in a project earlier today
And I was making some basic changes and requesting different features or whatever and it just understood things so much better than Claud Sonic 3.5 like it was it's noticeable like it's very noticeable so more and more different tools are incorporating 3.7 you can see perplexity here did so so all in all guys that's pretty much it for this video I wanted to give you guys an overview of Claud Sonet 3.7 talk about some of the benchmarks as well show you a
Couple examples and like I mentioned I'm going to be doing more videos on this channel about 3.7 using an incursor building actual projects out showing you some applicable examples on how to use it in real world scenarios so if you want to stay up to date on all that as well as the latest AI coding AI agents Marketing sales business growth all that good stuff make sure to smash that subscribe button to stay up to date with the videos and then
Also too if you got some value here make sure to drop a like and comment down below what your thoughts are about 3.7 how do you think in compares to any of these other models how do you think opening ey is going to retaliate with GPT 4.5 and what have you been using Clawson 3.7 to build I'm very curious let me know what your thoughts are in the comments down below if you've noticed any strengths it has any weaknesses it has
One thing I'll say is that when it comes to generating many different you know lines of code or many different files it is absolutely unmatched I've never seen a model generate that many lines of code but other than that guys if you haven't already joined our free Fook group and Discord Channel stri community.com I'll leave a link down below for that and then also too guys definitely check out our stride AI Academy for behind the scenes resources tools trainings and more
And then also too guys if you run a business and you need help implementing AI agents or AI custom grow Solutions into your business we a call down below at executives.com and we can see if it's a fit or not also too guys if you're an entrepreneur founder Dev whatever the case may be and you want to actually sell these same solutions to other business owners any you want our exact protocol on how to do this and charge 2 to 10K
Plus book a call down below as well other than that guys I will see you in the next video keep hustling keep grinding and of course guys accelerate your stride take care are you tired of pouring thousands of dollars into appointment Setters only to watch watch leads slip away imagine having a team of elite sales agents booking qualified appointments for you around the clock no more wasted time on training no more frustration with performance and no more draining your budget on
Inconsistent and expensive call centers introducing stride agents AI powered appointment seters that work 24/7 never get tired and book appointments while you sleep trained on thousands of successful conversations our AI agents outperform human teams at just on10th of the cost join the ranks of businesses that doubled their appointments and booking rates in just a matter of weeks don't get left behind in the AI Revolution visit stride agents.com now and transform your entire sales process with Cutting Edge AI technology it's time
To accelerate your stride with AI agents