GLM 4.7 vs MiniMax-M2.1: Best FREE Open-Source AI Model?

GLM 4. 7 versus MiniAX 2. 1, which is the best free opensource AI models. So, these are two of the latest models that are absolutely dominating when it comes to open source. You can see they just both released this month. And you've got MAX 2. 1 from Miniax AI and you've got GLM 4. 7 from Zai. We'll be putting them to the test, testing them out. I'll show you how you can use them today for free

as well. And you can see some of the key benchmarks here. So for example, if we check this out, you can actually go to the benchmarks here and you can see how these models perform side by side, right? So for example, Gemini Pro, which is up here versus Miniax and Miniax 2. 1 is actually winning on a lot of benchmarks. If you want to see the benchmarks, for example, GLM 4. 7, let me show you how this works. So you can see here, for example, we've got GLM 4. 7 at the top, and it is dominating on benchmarks as well. Now, I always take these benchmarks with a pinch of salt, and you want to test these out yourself, which is what we're going to do today to see how they perform. Now, you can get

access to both of these models for free. The way that you would access these models for free, I'm going to show you how, is that you can actually use the inference model inside Miniax 2. 1 as here, right? So, if you go to this section, you can use GLM 4. 7 in the chat and just ask it prompts and you can use it for free there. You can also, for example, go to Miniax and they have a sandboxed AI chat here. Top of that, what you can actually do is you can go over to Miniax and use it locally from GitHub. So, you can see it's a fully open- source model. And additionally, both of these models are available on O Lama for free. So, if you download OAMA for free, it's free to download. Then go to models here and then inside this section, if you want to use Miniax for free, you can just type in Miniax here and you'll get access to Miniax on the cloud version for free. And there's just a small usage limit, but it's good enough to get results that have never hit the limit. And you can see examples here. You can also, for example, type in GLM 4. 7 and you'll get GLM 4. 7 that you can then pull into. So let me show you how this works. So you would just copy this, paste it into your terminal with running already, and then you can start using this directly. Okay, so that's basically how you can get access to both of these models for free using OAM. And we get straight into this. Now, if you want all the video notes from today's inside the AS success lab link in the classroom, just go to December 2025 link in the comments description if you want to get access to this stuff. We're going

to be comparing these. All right, so both of these models are from China. GLM 4. 7 is from China's Z AI. Miniax 2. 1 is from China's Miniax and both are free to use as I've already shown you today and both claim to actually beat Claude and GPT on coding. Both have 200,000 context windows. But which one is actually better? We'll be testing them live today and running them head-to-head. So, let's get straight into this. We're going to take a simple example like this and we can open up both of these chats like and we can go into GLM 4. 7 and plug in this prompt here. Now, you can select between deep think or non deep think right. So deep think you're going to get better results but it's going to be slower and you can for example type it in. So you can also select design over here and then you can say okay build a simple landing page for this website right you can do the same inside minimax inside the AI agent and then hit run like that and that's how you can compare both these models and see them side by side in terms of how they perform. All right so you can see this is beginning to code out in the canvas now and we also have deep thinking on this task as well. The other way that you can do this is you can actually go over to open router and compare them side by side like so. If we go to open routting go to models we'll type in GLM 4. 7 and then we can click on chat and click on plus and then compare like that and we can compare these side by side with the same prompt. So if we for example click on create landing page and we'll say build a simple landing page for an SEO agency and we'll go from there and you can see these are both coding out. So one thing that I'll say is like if you use the agent directly here you will get better outputers with miniax but it just takes a bit longer. It's a bit more timeconuming to do. If you go inside GLM 4. 7, we don't have deep think tagged on. But you can see how we can use this quite quickly and it's created this beautiful landing page that actually looks pretty impressive, right? For an open source model, it's not bad at all. And what's interesting here is like if you have a look, so this is the output from Miniax so far, right? It's completed the project deployment. We can open this up. So if we open up this website, this is the page from Miniax and GLM 4. 7. Right? What's interesting about this is how similar do they look, right? They've even, for example, used very similar copy and that sort of thing, right, in the actual content. So if we scroll down here, which one would I say is better? I would actually say the GLM 4. 7 created something nicer because it's actually created, for example, like a logo. There's a bit more content on the page, sharing results, etc. Both of them are pretty basic, but both of them look nice and clean as well. And you could always adjust the prompt to make them better. But I would say for that one, I'm going to go with GLM 4. 7. Let's test the next thing now. And you can see them side by side inside here as well. So this is inside Open Router. You've got the output from Miniax. Actually, to be fair, inside Miniax of Open Router, it looks a lot nicer than if you're using GL 4. 7. You can see how similar both of them code, right? They're both very similar in the outputs they give. So now

what we're going to do is we're going to test a new thing. So, we a new chat here. Run them both. And we'll say, okay, create a simple racing game with keyboard controls. Now, I want to switch to interactive app. Got MiniAS and GLM about to code out right there. We'll create the artifact and see which one comes up with the best option. And if you have any questions about GLM 4. 7 or MiniAX as we go along, feel free to ask. We can also compare them side by side here. So, if we go to open router, go to models, and whilst those models are coding out, let's compare them. So, we've got GLM 4. 7. We'll click on compare. And over here we're going to add a model and we will compare that to Miniax 2. 1. Let's grab those and we can see them side by side right here. So for example, here's how they perform on the benchmarks. They both have very similar context window which is just slightly above Claude's which I think is quite funny. So I think it's 200k context window. These are both over 200. So 203 and 205. They both have reasoning models. GLM 4. 7 has a lot more providers. Both of them are open source. free to use on O Lama. And you can see how they perform right there. You can also compare the cost here. So, for example, GLM 4. 7 is just slightly more expensive than miniax 2. 1. Minia 2. 1 is 30 cents per million tokens and GLM 4. 7 is 40 cents per tokens. So, it's 33% more expensive actually, which is quite a bit. And if we go back to open router and we'll see how they're coding out now. So, we've got both of these generating. So, I think these are still coding out. Let's go to preview and see what we've got. We have the neon racer directly from GLM 4. 7 here. So, let's test it out. And there we go. Look at that. That's pretty cool. That's quite an addictive game. So, you can see how the Neon Racer works. We're score racking up points there if we crash. There we go. And actually tracks a high score as well, but the game is fully functional and actually playable, which is pretty cool. This is from GLM 4. 7. If we have a look at Miniaax, it did fail on the preview here as you can see. And we'll start a new chat and set up a new test. In the meantime, we have them working inside the chat directly here as well. So, you are going to get different outputs inside the agent itself versus using Open Router if that makes sense. because inside open router is a bit more basic when you're using the chat and the API directly. But if I had to compare them side by side, I'm going to say that GLM 4. 71 again. All right, really powerful. I actually saw on X, let me find this. The GLM 4. 7 is featured on artificial analysis intelligent index and positioned as a leading open source model. Right? So you can see how it's performing here. It's right up there with GP 5. 1, Claude, etc. But the difference is these are all private models, right? Whereas this is an open source model and it's just beating for example Kimmy and some of these other models over here. So let's go back to open the task status update here. So this is Miniax now. This is fully deployed. The thing that I love about Miniax is like it deploys it automatically in the background to a custom domain. And like I said you do get better outputs inside the agent versus using open router. So if we click on here we can use this. It is always tough. That is a tough one. All right let's see if we can play this. Have to switch over quick. I'm locked in here. I'm locked in. That's pretty fun. I actually say that's a better game than the one from GLM 4. 7 over here. But we had to use the agent to generate it. So that's one good tip to learn there. If we have a look inside Z chat that is still building out using the thinking mode to really come up with a good game there. I'm actually going to go with if we use an agent, I'm actually going to say that Miniax won. Let me know in the chat which one you think was the best so far. But I do think both of these models very impressive what they've built so far. So I'm going to run up a new task here. We're going to now say build a

rock paper scissors game with animations inside Miniax and we'll do it inside open to chat to compare them both. This is super addictive. I will close that. If there's anything you want me to test as well for both of these models, just let me know in the chat. This is still generating. So max which has coded this out now. Let's go on the preview. Load that up. Now what you can actually see inside the chat when you're using the agent is that yes, it's a bit slow but actually tests the tool as well, right? So I would say it's probably better for just getting better outputs. Z chat 2 that seems to have failed on the preview inside open router. So, if we ex off the preview, I don't think that's very good. Let's have a look at GLM 4. 7. See if that managed to do something. Here we go. It's a draw. Oh, look at that. It's weird the layout, but you can play this rock paper scissors game and then it counts the scores up and it actually functions and works. So, I think the API is actually seems to be better inside Open Rift GLM 4. 7. Let's see what we get back inside the chats now. So, we've got the rock paper scissors arena and again like how much better does that look when you're using the agent versus using open rout directly. So, let's click on scissors now. See what we get. It's a tie. Hey, there we go. We win one point. I like the little confetti animation they did there as well. Oh, yeah. We're destroying the computer here on rock paper scissors, but it actually works. Actually functions nicely done. GLM 4. 7 seems to be a lot slower when doing this stuff. Prof says, "Bro, for building SAS or apps, which is better, Miniax or anti-gravity or claw? " Honestly, I'm going to go I would still use Claude. Out of all those models, I still prefer Claude. Anti-gravity is quite good, but it's hard to connect to back end and do it full stack. Build a 3D car game. I think we just did right with the game that we built out. I guess it was 2D though. So, we can build a 3D car game. Build a 3D car racer game. Z chat. Super slow on that. I'll leave that running in the background. Bear in mind one thing you got is like on mini map you get very limited credits on the free plan. I'm on the pay plan inside here whereas you can use it for free. With GLM 4. 7, it seems to be pretty much free. I haven't hit any limits yet so far. I'm sure I will hit limits, but I've not done it so far. So, we're going to say build a 3D racer car game there as well. No, GLM 4. 7 doesn't have limits. Wow, that is amazing. GLM 4. 7 is the BM. So, building

that 3D racer game with Miniax and GLM 4. 7. Now, this looks identical to Mini. Let's play this out. The only difference is that inside this one, it actually tracks the scores. So, it tracks the scores whereas GLM 4. 7 doesn't. And for that reason, we'll go with Miniax 2. 1. And so far it's pretty close, right? GLM 4. 71 on the landing page. Mini max one on the racer game. Rock paper miniax 2. 11. We're waiting for the 3D car racer game. In the meantime, we'll set up the new test. So, we've got the 3D racer car game here. There we go. Not bad. It's beginning to be built out as you can see. Not bad at all. All right. Let's have a look at this one. So, this is Neon Drift. This is the one from Miniax. This looks a bit more interesting. Let's see what we got here. That's pretty cool. Wow. It's a little bit buggy, but as long as you stay inside the This is crazy. That actually hurt in my eyes. All right. I would say that is a much more complex and interesting game. And that was the one from Miniax. It does track the score at the bottom. Let's have a look. Let's see. Ah, it does. Look at that. All right. We're going to put that as a draw then for rock paper scissors, ironically. All right. Thank you for pointing that out. Appreciate you. So, we got two more tests to run out now. We got the

whack-a-ole game. That game with mini maps was pretty amazing to be fair. So, we're going to build a fitness tracker dashboard with charts. Here we go. Run these two. See which one is the best. All right. We got the whack-a-ole game being polished out right here. Is a quick summary of what has happened so far here. Yeah. So, if you want to see the test so far, we're comparing GLM 4. 7 versus Miniax. And so far on the first four tests, GLM won the landing page test. Miniax won the racer game. It was a draw for rock paper scissors. Miniax won the 3D racer car game. And we're testing out the last three. Now, if you want all the video notes from today, then you can get that inside the AI success lab, which basically summarizes everything you need to know about all of these tools that we've covered today. So, for example, if you go inside the community here, link in the comments description or just go and chat GP or Google and ask for the AI success lab, you see we've got 40,00 members. If you go inside the classroom here, you can find all the video notes of all the tools that we cover, right? All the video notes and the links and the resources are inside here along with the notes and the video resources to Ger 4. 7 and Miniax 2. 1. So if you go inside here, we've got the whack-a-ole game ready to play. Let's play this out. Here we go. Start the game. Boom. This is crazy how good it is to be fair. Oh, it's too fast. I need my mouse. I'm using the trackpad. All right, that's insane but good. Very addictive. It's actually hard to click off that. Look, now we've got GLM 4. 7. I consider myself quite fast. Ah, it's such a good game. It's overwhelming, but it's great. Finally picked you as a mentor out of all the overwhelm. Thank you very much, sir. I appreciate you for watching and I'm glad that it helps you. That was hard. That was really hard. Do you know what? If you want to play this game, I'm going to share this link with you. I should grab all the links from today and actually share that with you, shouldn't I? Let's go back to Max now. So, if you want to see the outputs from today for Miniax, cuz it deploys it on a open web domain. And I would actually say that just won on that one simply because it created points that pop up when you actually whack the mole, right? But we'll come on to that in a second. So, if you want to get all the links from the games, you want to play these for yourself, you're inside the AI success lab. And then I'll put the games links here as you can see. All right. And you can get access to all those for free as

well. I think I'm going to stop it there cuz that was overwhelming. But we've done five tests and Miniax I think crushed it if you have a look here. So, we've done five tests. We did the landing page, the racer, rock paper scissors, build a racer car game, and also build a whack-a-ole clicking style game. GLM 4. 7 won the landing page. It was a draw on rock paper scissors, but Miniax crushed it on the racer game, the 3D racer car game, and build a whack-a-ole game as well. The only difference here is that obviously GLM 4. 7 is free to use inside the chat. So, if you want a free option, just go with 4. 7. Like the differences between them are very minimal, but if you want something that's like the ultimate agent, and I've been very impressed with Miniax. The more I use it, the more I like it, actually, to be fair. And you can see even here like the I love the fact that it just it automatically deploys this stuff, right? So here's the fitness tracker app. As you can see, it's created a tracker. We can add stuff to it. We switch between month, year, etc. We got the steps and everything else. Nice little dashboard right there. You can build stuff out and just deploy it onto a subdomain. It feels a bit like Manners of Gen Spark, but faster and better. I'm really loving these tools. I think it's a great super agent. If you again, if you want the premium version and you don't really care about costs, then go with Miniax. If you want a free version and you don't want to hit limits, then go with GLM 4. 7. All right

so thanks so much for watching. Appreciate everyone who watched today. As you can see, you know who the winner is. We've been through all the tests. I've given you access to all the games that we built from today inside the AI success lab along with all the links for all these tools. So, for example, the hugging face, the free sandbox, open router, the GitHub, the API docs, etc. It's all inside there. And additionally, we've compared all the benchmarks. And also I've actually done is if you want to use either of these tools, I've actually created a 30day open-source AI domination plan that shows you each day what to implement to get results with this stuff using the custom framework that I run. We also giving you a bunch of prompts for like head-to-head stuff that you can test. And also we have created a list of 100 prompts that you can test this stuff out with, right? So 100 prompts for comparing GLM 4. 7 versus Miniax. As you can see right here, Prof says, "Love you, bro. Thank you very much, man. I appreciate you and yeah, I appreciate you watching. Now, you know, AI is changing overnight. It's changing business especially. Whilst most people are still figuring out basic prompts, AI profit boardroom members are building real systems that automate the work. If you haven't already, check out the AI profit boardroom link in the comments description or go to the AI profitboarding. com. You can see an example what it looks like right here. If you go to the community, you can see that we actually have a daily accountability group. So, you can jump in the group and post your goals, etc. You can also go inside the community, ask questions, get help, get support whenever you need it. And additionally, what we have over here is we have a win section. So you can post your wins, share tactics, share what you're learning and building, which is pretty amazing. You can see, for example, like Vladimir is building mini courses. You've got, for example, Johan working on white labelled voice agents. You've got, for example, people building their own homegrown MCPs. even daily doses of inspiration from Steve Light and NA10 automations with Stephen who he came into the group and he didn't even know anything about NA10 and now he's absolutely crushing it and he's built automations that have saved him like 10% of his time. So if you want to get all my best trainings an amazing community four weekly coaching course you can get that all inside the AI profit boardroom link in the comments description. I appreciate you watching today and I'll see you on the next one. Cheers everyone. See you. Bye-bye.