AI News Today
← All episodes
Episode 35 · June 16, 2026 · 13:43

China’s GLM 5.2 is TERRIFYING!

GLM 5.2 vs Claude Opus 4.8: I Ran 5 Build Tests (China Won 4) + The Catch

The video compares China’s new GLM 5.2 (by ZAI, released June 13, 2026) against Anthropic’s Claude Opus 4.8 (released May 28) using the same five build prompts. GLM 5.2 produces better results in four tests—a running game, an Apple-style landing page, a liquid-sloshing animation, and a neon arcade game—while Claude wins the fifth test with a cleaner moving solar system map. The script highlights pricing differences (Claude’s per-token API costs vs GLM’s flat monthly coding plan) and argues the best approach is to run both: use GLM for cheap “building” tasks and Claude for deeper thinking inside a unified “agent operating system.” It notes key catches: no independent benchmarks yet, and GLM’s API/chatbot/open weights are coming next week.

00:00 GLM vs Claude Setup
01:28 Build Test 1 Runner Game
02:01 Build Test 2 Landing Page
02:26 Build Tests 3 and 4 Animations
03:48 Test 5 and Price Shock
04:54 Stop Picking Sides
05:26 Agent OS Workflow
07:01 The Catch and Missing Proof
08:14 Agent OS Offer Break
09:13 Daily Business Use Example
11:13 Why Agents Beat Model Chasing
12:28 Final Verdict Run Both

Full transcript

China's GLM destroys Claude that's the claim flying around right now so instead of guessing I ran the same tests myself same prompts same jobs GLM 5.2 from China on one side and then we have Claude Opus 4.8 from America on the other side and here's what happened on four out of five build tests the Chinese model made the better thing the more fun game the smoother page I'm going to show you exactly what we're built today with it and how it works and how this outperforms Claude 4.8 on many of the tests so the crazy thing about this is Claude actually lost most of the tests you might be thinking now hold on you know before you think this is some China good America bad video it's not it's way more useful than that because by the end you'll know which one to use when to use it and the smarter move that beats picking sides at all and there's a catch with GLM 5.2 that almost nobody is talking about so I'll get to that it changes everything but first the tests let me set the scene GLM just came out on June the 13th 2026 it's built by a Chinese company called ZAI used to be Jerpo and Claude Opus 4.8 is built by Anthropic in America came out May the 28th both brand new both top tier so I gave them the same five jobs and just watched the first test was a running game a simple running game just dodge blocks grab coins speed picks up as you go and GLM built the fun one you can see it right here so this is the test and we use the same prompt on all of these but this was the most fun one that we actually created and you can see how cool this is we actually tested out uh Kimi K215 which came nowhere near but here's the crazy thing Claude actually lost I mean look how boring this game is compared to the other one GLM 5.2 actually beat it and it would create a way more fun situation second test a landing page so we've got GLM 5.2 here we have Opus 4.8 here I asked both for a clean Apple style launch page for a made up product GLM 5.2 absolutely nailed it let me show you this it looked premium the menu words the scroll felt nice Claude's page had less on it and actually felt flat when we tested out this is Claude's page right here it just feels a little bit flat a little bit boring in comparison to GLM 5.2's version third test liquid sloshing in a bowl and Kimi K217 and Claude Opus made something really boring here but look at the quality of GLM 5.2 this is GLM 5.2's output you can change the theme you can make it more interesting it just looks way cooler sounds a bit silly to test but it's a real test of how well the model handles tricky animation GLM 5.2 made the most fun one you could change the colors it feels alive Claude's actually fades out really fast when you try this out seems super boring kind of works but you can only change two colors and they're both really boring so China won round three and then there was a fourth test a neon arcade game so you can see that right here now if we have a look at GLM 5.2's version it built something genuinely fun and wild look how cool that is right that's quite a cool fun game to play and genuinely fun Claude's worked but it feels a little bit buggy when you use it and also it's just a little bit boring if you have a look at this like this is pretty standard run-of-the-mill stuff compared to GLM 5.2 so China actually won round four that's four wins for China's AI model GLM 5.2 now here's where I have to be straight with you because I'm not here to hype Claude didn't lose everything on the fifth test a moving map of the solar system Claude Opus 4.8 actually made the best version cleaner better looking so it's not a clean sweep Claude is still excellent at certain jobs but the pattern is hard to ignore a model from China that came out this week literally just dropped within the last 24 hours beat one of the best American models on most of the things that I threw at it and here's a part that should make you sit up as well is the price Claude Opus 4.8 costs six sorry five dollars per million words in 25 dollars per million words out that adds up fast if you're using it all day on the API GLM 5.2 runs on a coding plan that starts around 18 dollars per month and that's not per use that's a whole month month including plugging into your AI agents now as of June the 13th 2026 GLM 5.2 is available to all users of the GLM coding plan across the light pro and max tiers so you've got a model from China beating Claude on most tests for a flat price that's a fraction of the cost and you see why people are losing it now let me break a belief you might be holding right now a lot of business owners think the cheap Chinese models must be junk knockoffs toys that belief is costing them money because the thing I just showed you wasn't a toy it built a better landing page than most of the expensive models on earth the old way was pay the most get the best the new way means that is not true anymore the new way is really just test them and use the right one for the job and that brings me to the real lesson here because picking aside China or America doesn't really make sense that is the wrong game let me show you what I mean this morning for example I didn't pick one I had GLM 5.2 building amazing stuff out inside the agent operating system and I had Claude doing the deep thinking work and had them both running inside one dashboard I call the agent operating system one screen all my agents talking to each other working together and that's a move nobody is talking about you you don't choose China or American models you run both you let the cheap one do the heavy lifting and the smart one handle the hard parts and you orchestrate them from one place think about what that means for your business you can have GLM 5.2 building all your landing pages for almost nothing whilst Claude handles your customer email so you can have both running at the same time that's the whole point of the agent OS it wires your agents together into one mission control your Claude your GLM your free agents all in one spot you're not jumping between 10 tabs you're not paying for one expensive model to do everything you give each job to the cheapest agent that can do it well and here's a key detail that most people are missing you can plug GLM 5.2 into your AI agents you can't do that as easily with Claude inside other agents with GLM 5.2 on that coding plan you just drop it into your setup and it just runs so we've already plugged it into Hermes agent for example now day one support with GLM 5.2 includes Claude code Klein open code root code goose crush open chlorine kilo code so your whole agent army can run on a flat monthly price instead of bleeding you per use and that's the unlock not which model runs or wins but how do I run them all for cheap in one single place now let me show you the catch I promised as well because this matters before you go all in on GLM 5.2 there are no official benchmarks not that I've seen anyway no independent benchmark results were published at launch the company says it's better than past GLM versions and long coding tasks but third-party checks are still pending this is brand new so when someone says GLM 5.2 beats Claude on the numbers they're guessing the numbers don't exist yet unless they've created their own benchmarks which again wouldn't be an independent test so what I showed you is real I've tested with my own eyes but that's my test not a lapse so I've got to be honest about that the hype online is running way ahead of the proof two more things that you should know as well the API the chatbot and the open weights are all coming next week not yet so for example if you actually go to ZAI you can't use it directly what I can say is it's built some amazing stuff so for example this video that you're watching right now was fully edited with our Hermes agents plugged into GLM 5.2 inside our agent operating system the open version the one you can run free on your own computer that isn't out yet it's prioritized and promised but not shipped so I'm telling you this because most videos won't they tell you it's a Claude killer and stop there the truth is more useful it's incredibly good at building things it's cheap but the proof is still coming and some of the best parts haven't even shipped yet quick thing before I show you the rest if you want the exact agent operating system I'm running right now with GLM 5.2 with Claude plugged in with Hermes and GLM 5.2 and everything else we've even got Hermes Jarvis which is a voice activated version of Hermes and this is all plugged in and talking to each other that's inside the AR platform you get the agent operating system as a zip file you can just install drop in your GLM 5.2 your Claude your free agents all-in-one mission control plus a 30-day plan on exactly how to implement all this stuff into your actual business and that walks you through wiring GLM 5.2 into your agent step by step so you're building landing pages and content for almost nothing instead of paying per use we've got four coaching calls every week inside there where you can share your screen and we'll help you set up live and inside the community loads of members are already running GLM and Claude together for client work content lead generation to help you 24 7 so link in the comments description or go to the airprofitbomb.com now back to it let me show you the smartest way to actually use this day to day in a real business here's how I run it so I have Claude desktop operating my agent operating system and just building stuff and wiring in there right so it hands jobs out to the other agents GLM 5.2 does the building the cheaper agents do the bulk work it all flows beautifully into the agent operating system where I can see everything and everything that I've created inside here beautifully and you can see how cool it is and how easy it is to just preview this stuff and see what we've created I think that's absolutely amazing and the perfect way to use it and so you can have like a live activity feed you can have each agent its own workspace a shared memory so they will remember the same stuff because we have obsidian plugged into this as well and so that last part is big like a shared memory it means your agents don't just start from scratch every time they remember your business your customers your brand your voice so the work gets better the longer you run it and here's a real example anyone can use you could have GLM 5.2 spin up five different landing pages for your offer overnight then have Claude pick the best one in the morning with one sentence you just replaced a week of work in a designer's um bill now let me write the second belief that's probably stopping you right now a lot of people think this agent stuff is just for coders I'm not technical for example or this isn't for me wrong dead wrong because inside the air profitable and we've actually helped over 3,500 business owners as you can see with all these testimonials 181 pages testimonials of people doing the exactly this and the load of them have never touched AI before they joined I'm not saying that to boast but I've seen agency owners freelancers shop owners not coders and they're running these agents right now to get more leads and get more customers so if they can do it so can you and here's the thing about timing as well we're early on all of this this literally just dropped today bear in mind it's pretty much the same day that Fable 5 got banned so they're still you know most people are still doing everything by hand and the ones who learn to run agents like this now whilst it's new get a head start that's really hard to catch and that's the whole game right now you know if you move early you get the advantage now let me zoom out and tell you what this race actually means for you because China is shipping fast like really fast ZAI shipped four flagship models uh in roughly four months that's a new top model every few weeks in America shipping just as fast you know Claude Opus 4.8 came out just 41 days after Opus 4.7 that's the shortest gap between Opus releases yet so what does that mean for you well it means the model you pick today probably won't be the best one next month there's always a newer cheaper better one coming if you build your whole business around one model you're probably stuck when the next one drops you're just starting over again that's the real reason the agent operating system matters because we can just swap models in and out if a new model comes out no problem we just plug it in you don't have to rebuild everything and whilst everyone is panicking about which model was best this week you just slot the new one in and keep going you stop chasing models you start running a system and let me break the last belief because it's the one that keeps most people stuck as well people think I'll wait until this improves or until there's a clear winner then I'll learn it but it's never going to settle down there's no finish line coming the people winning aren't the ones who waited they're the ones who started learning the system whilst it was messy so that every model just makes them stronger and better waiting isn't safe waiting is how you fall behind so there's where we land so did China's GLM 5.2 destroy Claude on most of my build tests yes it made the better thing for a fraction of the price that's real and it's a big deal but the smart play isn't to ditch Claude and run to China the smart pet play is to run both let GLM 5.2 build cheap and fast let Claude handle the hard thinking and together you can tie them so that they work as one team whilst you sleep 24 7 and that's not picking a side that's building a machine that uses whatever's best whenever for almost anything if you want that exact machine the agent operating system I'm running with GLM 5.2 and Claude both plugged in it's inside the AR profit boardroom you get the full agent operating zip file to install the 30-day roadmap for wiring GLM 5.2 into your agents to pump out landing pages and content cheaply and automate whatever you want you get four weekly coaching calls where we set up with you live daily tutorials including how to use GLM with Hermes and Claude together and three thousand six hundred business owners who are already running these agents to get more leads and more customers ready to help you at any time we update agent operating systems as new models like GLM 5.2 drop as well so you're always running the latest link in the description or go to the arprofitboard.com test them both run them together and start now whilst it's early that is the move

More episodes

Browse all episodes →