Loading video...

Video Failed to Load

There was a problem loading this video. This could be due to a temporary network issue or the video might be unavailable.

Today, we're introducing GLM-5.2 Fast: our GLM-5.2 Model API designed for the most demanding real-time use cases. The Fast tier delivers 2-3x higher TPS than the standard GLM-5.2 MAPI.

Baseten

23,215 subscribers

115,626 views • 3 months ago •via X (Twitter)

Science & Technology

Anya Rossi• Live Now

Private livecam show

25 Comments

Baseten3 months ago

More in our blog: Try it now:

Mia3 months ago

This is insane 🔥

Jernej Strasner3 months ago

Not as fast as @sferenceai though 🙃

JesseEvans3 months ago

Is it the vision version?

Quantinine3 months ago

this is awesome. Will you be supporting kimi k3 release this weekend?

Baseten3 months ago

Absolutely

Eitan Turok3 months ago

what is in/out toks/sec?

sofia3 months ago

so does "fast" mean faster tokens or faster latency

Denis Loginoff ⚡️3 months ago

Does it have Zero Data Retention btw 👀

Launch Archive3 months ago

GLM-5.2 Fast API offers 2-3x higher transactions per second for real-time applications, enhancing performance for users. Congrats @baseten on #1 on Launch Archive today.

Jaume Martin3 months ago

It's good, but... I hit this wall constantly... :(

KimSunIl3 months ago

총 3 x DGX Spark 로 실행하는거니? glm5.2 는 어떤 모델인가요

Sebastian Buzdugan3 months ago

2-3x tps helps but p99 under bursty load decides real-time fit

Stephane3 months ago

Nice. Can you do the same with KimiK3?

Madison Kanna3 months ago

🥳🥳🥳

vinayak 🦕3 months ago

Need

felix3 months ago

That's impressive, if it can maintain comparable performance //'GLM 5.2 Fast serves the same weights as standard GLM-5.2'

AI Mastery Guide3 months ago

2-3x higher TPS is a big deal for anything running in real time.

CV.YH3 months ago

amazing work

武止戈👽🦀相比于《1984》, 我宁可《2012》3 months ago

我需要“slow” api，只要不慢10倍就行，我可以用并行subagents来确保进度不被显著拖慢，只要成本能降下来就行

Jurly3 months ago

2-3x TPS improvement is significant, what tradeoffs did you make in accuracy or context retention?

pccallout3 months ago

Price ?

Tushar Koshti3 months ago

#glm fast is really helpful interns of speed

lynn tam3 months ago

price?

x863 months ago

Do that with k3 soon and you might just win

Related Videos

GLM 5.2 runs pretty fast on Modal.

GLM 5.2 runs pretty fast on Modal.

Charles 🎉 Frye

20,342 views • 4 months ago

This is GLM-5.2 running on 3x DGX Sparks

This is GLM-5.2 running on 3x DGX Sparks

Mia

41,809 views • 3 months ago

GLM 5.2 Easy Setup with Baseten + Droid harness. 00:00 Fastest GLM 5.2 Setup Goal 00:32 Getting a Baseten API Key 01:09 Automated Droid AI Configuration 02:23 Testing the GLM 5.2 Integration 03:07 Alternative Setups and Limitations 04:14 Bonus: Setup for Other IDEs

GLM 5.2 Easy Setup with Baseten + Droid harness. 00:00 Fastest GLM 5.2 Setup Goal 00:32 Getting a Baseten API Key 01:09 Automated Droid AI Configuration 02:23 Testing the GLM 5.2 Integration 03:07 Alternative Setups and Limitations 04:14 Bonus: Setup for Other IDEs

Ray Fernando

10,580 views • 4 months ago

GLM-5.2 delivers a substantial leap in app development capabilities, which also represent demanding long-horizon tasks. Results: - GLM-5.1: 21/70 - GLM-5.2: 48/70 - Claude Fable 5: 56/70 That's more than a twofold improvement from GLM-5.1 to GLM-5.2. These come from an internal benchmark of 35 challenging mobile development tasks, each run twice for a total of 70 trials. We measured task completion, defined as core features working without major issues.

GLM-5.2 delivers a substantial leap in app development capabilities, which also represent demanding long-horizon tasks. Results: - GLM-5.1: 21/70 - GLM-5.2: 48/70 - Claude Fable 5: 56/70 That's more than a twofold improvement from GLM-5.1 to GLM-5.2. These come from an internal benchmark of 35 challenging mobile development tasks, each run twice for a total of 70 trials. We measured task completion, defined as core features working without major issues.

Zixuan Li

346,910 views • 4 months ago

GLM-5.2 is a bigger deal than you think. > what GLM-5.2 actually scores (0:46) > "it's just a rich person's toy" (4:27) > why GLM-5.2 matters even if you never run it (7:19)

GLM-5.2 is a bigger deal than you think. > what GLM-5.2 actually scores (0:46) > "it's just a rich person's toy" (4:27) > why GLM-5.2 matters even if you never run it (7:19)

Rob Hallam

19,906 views • 4 months ago

we asked GLM 5.2 and Claude Opus 4.8 to find and fix the bugs in two different scripts while Opus 4.8 finished a few seconds faster in both tests, GLM 5.2 burned roughly 17k–38k fewer tokens and absolutely crushed it on cost, coming in 15x to 22x cheaper! Test 1: GLM 5.2 ($0.068) vs. Opus 4.8 ($1.06) Test 2: GLM 5.2 ($0.045) vs. Opus 4.8 ($1.02)

we asked GLM 5.2 and Claude Opus 4.8 to find and fix the bugs in two different scripts while Opus 4.8 finished a few seconds faster in both tests, GLM 5.2 burned roughly 17k–38k fewer tokens and absolutely crushed it on cost, coming in 15x to 22x cheaper! Test 1: GLM 5.2 ($0.068) vs. Opus 4.8 ($1.06) Test 2: GLM 5.2 ($0.045) vs. Opus 4.8 ($1.02)

GMI Cloud

86,766 views • 4 months ago

GLM-5.2 is now available on AI/ML API! We asked it and Opus 4.8 to one-shot a playable Backrooms, same prompt. Opus got the flashlight working but you can't run and can't pause. GLM-5.2 built the full mechanics. Opus: 2:14sec & $1.94 GLM: 1:08sec & $0.37 Docs, guides & setup below:

GLM-5.2 is now available on AI/ML API! We asked it and Opus 4.8 to one-shot a playable Backrooms, same prompt. Opus got the flashlight working but you can't run and can't pause. GLM-5.2 built the full mechanics. Opus: 2:14sec & $1.94 GLM: 1:08sec & $0.37 Docs, guides & setup below:

AI/ML API

475,535 views • 4 months ago

GLM-5.2确实有点东西

GLM-5.2确实有点东西

关木

58,366 views • 5 months ago

you can now get unlimited free GLM 5.2 for the first time in history 😳 this is bigger than ever. here's how to do it: 1 / cancel your existing subscriptions: cancel your Lovable, Cursor, and Claude Code subscriptions. you don't need them anymore 2 / sign up for an account on freebuff you can now use unlimited GPT 5.6 and deepseek as well. 3 / to get GLM 5.2: go to the earn tab, and complete the bounties. a bounty provides unlimited GLM 5.2 messages for an hour, each. this isn't going to last forever.

you can now get unlimited free GLM 5.2 for the first time in history 😳 this is bigger than ever. here's how to do it: 1 / cancel your existing subscriptions: cancel your Lovable, Cursor, and Claude Code subscriptions. you don't need them anymore 2 / sign up for an account on freebuff you can now use unlimited GPT 5.6 and deepseek as well. 3 / to get GLM 5.2: go to the earn tab, and complete the bounties. a bounty provides unlimited GLM 5.2 messages for an hour, each. this isn't going to last forever.

Victor Cheng

55,221 views • 3 months ago

Looking like we'll 4x revenue in the next 90 days, thnx GLM 5.2

Looking like we'll 4x revenue in the next 90 days, thnx GLM 5.2

Sam Hogan 🇺🇸

26,460 views • 4 months ago

I was skeptical about giving GLM-5.2 the task to build a super mario snes style game in a single HTML file so i ran the test on Opus and also on GLM-5.2 Opus didnt even try to make a game or made something but it wasn't a game GLM-5.2 one shotted a pretty functional game.. WILD

I was skeptical about giving GLM-5.2 the task to build a super mario snes style game in a single HTML file so i ran the test on Opus and also on GLM-5.2 Opus didnt even try to make a game or made something but it wasn't a game GLM-5.2 one shotted a pretty functional game.. WILD

Hermes Agent Tips

32,883 views • 4 months ago

Grok 4.5 operates at Opus-level intelligence and Sonnet-level pricing! We tested it against Claude Opus 4.8, Sonnet 5, and GLM 5.2 on a SpaceX mission: liftoff, stage sep, orbit insertion, satellite deploy. Grok 4.5 animates it more accurately and vividly than Sonnet 5 and GLM 5.2. Opus is the fastest and most expensive. Grok burns more tokens but still lands ~40% cheaper than Opus, and a bit under Sonnet 5. Time / tokens / cost: Grok 4.5 — / 35k / $0.69 Opus 4.8 — / 13.7k / $1.10 Sonnet 5 — / 22.1k / $0.78 GLM 5.2 — / 92.7k / $0.13 yet GLM did the whole mission for $0.13 😮

Grok 4.5 operates at Opus-level intelligence and Sonnet-level pricing! We tested it against Claude Opus 4.8, Sonnet 5, and GLM 5.2 on a SpaceX mission: liftoff, stage sep, orbit insertion, satellite deploy. Grok 4.5 animates it more accurately and vividly than Sonnet 5 and GLM 5.2. Opus is the fastest and most expensive. Grok burns more tokens but still lands ~40% cheaper than Opus, and a bit under Sonnet 5. Time / tokens / cost: Grok 4.5 — / 35k / $0.69 Opus 4.8 — / 13.7k / $1.10 Sonnet 5 — / 22.1k / $0.78 GLM 5.2 — / 92.7k / $0.13 yet GLM did the whole mission for $0.13 😮

GMI Cloud

151,027 views • 4 months ago

I just tested GLM 5.2, Claude opus 4.8, and GPT 5.5 to see which one could rebuild my personal site in that whimsical ascii style GLM 5.2 held its own and went toe to toe with the other 2 top tier models lowkey impressed, look at the hero section it built here

I just tested GLM 5.2, Claude opus 4.8, and GPT 5.5 to see which one could rebuild my personal site in that whimsical ascii style GLM 5.2 held its own and went toe to toe with the other 2 top tier models lowkey impressed, look at the hero section it built here

Ann Nguyen

33,873 views • 4 months ago

Nvidia could release a GLM 5.2 or better equivalent model any time it wants, and nuke OpenAI's margins 🤯

Nvidia could release a GLM 5.2 or better equivalent model any time it wants, and nuke OpenAI's margins 🤯

Marcelo P. Lima

102,723 views • 4 months ago

Coding Benchmark: Cursor Composer 2.5 vs Z.ai GLM 5.2 Task: DLA (Diffusion-Limited Aggregation) Judge: GPT 5.5 Pro 🎉Winner: Composer 2.5🎉 Sebep: Prompt’un “crux” dediği nokta performanstı: en az ~50.000 particle’da responsive kalma, optimize algoritma, live stats ve performans barı. Composer çözümü bu gereksinime daha doğrudan oynuyor: sabit 720x720 lattice grid, Int32Array, Int16Array, offscreen ImageData, MAX_WALKERS = 192, adaptive performance loop ve gerçek bir performance bar var. Prompt açıkça 50k real-time performansı ve box-counting doğrulamasını kritik görüyor. Final Puantaj ve Özet: - Cursor Composer 2.5 - 83/100 - GLM 5.2 - 80/100 GLM 5.2 zayıflıkları - Performance bar yok. Prompt bunu açıkça istiyordu. - Seeded determinism yok. Math.random() doğrudan kullanılıyor. - boundingR güncellemesi gecikmeli. Spawn bölgesi hızlı büyümede stale kalabilir. - Map tabanlı grid 50k+ için typed-array grid kadar stabil değil. - Start/Pause tek toggle. Kullanılır ama Composer kadar net değil. Cursor Composer 2.5 zayıflıkları - Sabit 720×720 dünya. Büyük aggregate ve uzun koşularda sınır etkisi üretir. - 4-neighbor lattice artifact riski. Daha kare/piksel tabanlı görünür. - Glow çok temel. Gerçek bloom değil, shadowBlur benzeri. - Seeded determinism yok. O da Math.random() kullanıyor. - Fraktal estetiği GLM kadar organik olmayabilir. 1)Prompt uyumu - Composer 2.5 : 8.6/10 - GLM 5.2 : 8.1/10 2)DLA algoritma doğruluğu - Composer 2.5 : 8.2 / 10 - GLM 5.2 : 8.0 / 10 3) Performans mimarisi - Composer 2.5 : 9.0/10 - GLM 5.2 : 8.0/10 Görsel kalite / seyir zevki - Composer 2.5 : 7.3/10 - GLM 5.2 : 8.4/10 UI / kontrol paneli - Composer 2.5 : 8.7/10 - GLM 5.2 : 8.2/10 Live stats / correctness check - Composer 2.5 : 8.8/10 - GLM 5.2 : 8.4/10 Ölçeklenebilirlik - Composer 2.5 : 7.5/10 - GLM 5.2 : 8.3/10 Kod mimarisi / okunabilirlik - Composer 2.5 : 8.4 / 10 - GLM 5.2 : 8.0/10 Risk / bug potansiyeli - Composer 2.5 : 7.8/10 - GLM 5.2 : 7.2/10

Coding Benchmark: Cursor Composer 2.5 vs Z.ai GLM 5.2 Task: DLA (Diffusion-Limited Aggregation) Judge: GPT 5.5 Pro 🎉Winner: Composer 2.5🎉 Sebep: Prompt’un “crux” dediği nokta performanstı: en az ~50.000 particle’da responsive kalma, optimize algoritma, live stats ve performans barı. Composer çözümü bu gereksinime daha doğrudan oynuyor: sabit 720x720 lattice grid, Int32Array, Int16Array, offscreen ImageData, MAX_WALKERS = 192, adaptive performance loop ve gerçek bir performance bar var. Prompt açıkça 50k real-time performansı ve box-counting doğrulamasını kritik görüyor. Final Puantaj ve Özet: - Cursor Composer 2.5 - 83/100 - GLM 5.2 - 80/100 GLM 5.2 zayıflıkları - Performance bar yok. Prompt bunu açıkça istiyordu. - Seeded determinism yok. Math.random() doğrudan kullanılıyor. - boundingR güncellemesi gecikmeli. Spawn bölgesi hızlı büyümede stale kalabilir. - Map tabanlı grid 50k+ için typed-array grid kadar stabil değil. - Start/Pause tek toggle. Kullanılır ama Composer kadar net değil. Cursor Composer 2.5 zayıflıkları - Sabit 720×720 dünya. Büyük aggregate ve uzun koşularda sınır etkisi üretir. - 4-neighbor lattice artifact riski. Daha kare/piksel tabanlı görünür. - Glow çok temel. Gerçek bloom değil, shadowBlur benzeri. - Seeded determinism yok. O da Math.random() kullanıyor. - Fraktal estetiği GLM kadar organik olmayabilir. 1)Prompt uyumu - Composer 2.5 : 8.6/10 - GLM 5.2 : 8.1/10 2)DLA algoritma doğruluğu - Composer 2.5 : 8.2 / 10 - GLM 5.2 : 8.0 / 10 3) Performans mimarisi - Composer 2.5 : 9.0/10 - GLM 5.2 : 8.0/10 Görsel kalite / seyir zevki - Composer 2.5 : 7.3/10 - GLM 5.2 : 8.4/10 UI / kontrol paneli - Composer 2.5 : 8.7/10 - GLM 5.2 : 8.2/10 Live stats / correctness check - Composer 2.5 : 8.8/10 - GLM 5.2 : 8.4/10 Ölçeklenebilirlik - Composer 2.5 : 7.5/10 - GLM 5.2 : 8.3/10 Kod mimarisi / okunabilirlik - Composer 2.5 : 8.4 / 10 - GLM 5.2 : 8.0/10 Risk / bug potansiyeli - Composer 2.5 : 7.8/10 - GLM 5.2 : 7.2/10

Alican Kiraz

17,562 views • 4 months ago

We ran GLM 5.2 and Kimi K3 on 100 deep research tasks (DRACO by perplexity) and asked Fable to be the judge Kimi K3: 71.6 mean score, 77% pass rate GLM 5.2: 41.5 mean score, 22% pass rate K3 took 3h 38m and $60. GLM finished in 1h 45m for $12.50. Deep Dive in 🧵

We ran GLM 5.2 and Kimi K3 on 100 deep research tasks (DRACO by perplexity) and asked Fable to be the judge Kimi K3: 71.6 mean score, 77% pass rate GLM 5.2: 41.5 mean score, 22% pass rate K3 took 3h 38m and $60. GLM finished in 1h 45m for $12.50. Deep Dive in 🧵

GMI Cloud

111,643 views • 3 months ago

I love using GLM 5.2 for web app iteration. My workflow: generate 6 variations, then pick the best one and continue iterating on it. I built Recast to make this even easier. Give it a prompt, get 6 variations, download the code for your favorite version, then continue iterating in your own coding agent. GLM 5.2 works really well for this use case since it's fast, cheap, and great at design. On average, you can make 3-6 landing pages with GLM 5.2 for every one you can generate with Opus 4.8. And the design variety of the GLM landing pages helps a lot with trying to think through what style you want to go with. Recast is launching soon (as a web app + skill).

I love using GLM 5.2 for web app iteration. My workflow: generate 6 variations, then pick the best one and continue iterating on it. I built Recast to make this even easier. Give it a prompt, get 6 variations, download the code for your favorite version, then continue iterating in your own coding agent. GLM 5.2 works really well for this use case since it's fast, cheap, and great at design. On average, you can make 3-6 landing pages with GLM 5.2 for every one you can generate with Opus 4.8. And the design variety of the GLM landing pages helps a lot with trying to think through what style you want to go with. Recast is launching soon (as a web app + skill).

Hassan

20,164 views • 4 months ago

We're proud to be the fastest inference provider on Artificial Analysis, OpenRouter, and Hugging Face for GLM-5.3-Flash, at 122+ TPS. All served from the US only, starting on day 0, with ZDR by default. Stay tuned for updates as our engineers continue to optimize GLM-5.3-Flash for even higher throughput and lower latency.

We're proud to be the fastest inference provider on Artificial Analysis, OpenRouter, and Hugging Face for GLM-5.3-Flash, at 122+ TPS. All served from the US only, starting on day 0, with ZDR by default. Stay tuned for updates as our engineers continue to optimize GLM-5.3-Flash for even higher throughput and lower latency.

Baseten

23,114 views • 2 months ago

Запуск ИИ на 744 МЛРД параметров БЕЗ видеокарты! Обзор GLM-5.2 локально

Запуск ИИ на 744 МЛРД параметров БЕЗ видеокарты! Обзор GLM-5.2 локально

Взаимная подписка с адекватными

11,172 views • 2 months ago

🚨 GLM 5.3 just dropped, and here’s how to access the GLM family completely FREE. no subscription. no API credit cost & no rate limit. step 1️⃣: download the OpenCode app on your PC. it’s available on both Mac and Windows. 🔗: step 2️⃣: go to NVIDIA’s website, sign up with your email, verify your account, and generate a unified API key for free. NVIDIA offers free model endpoints for developers, and GLM 5.2 is one of the top models currently available there. 🔗: step 3️⃣: launch OpenCode, open the model selection menu, search for NVIDIA, and paste the API key you generated. also, there’s no rate limit, but traffic to the model through NVIDIA can get busy at times, so responses may occasionally be slower or temporarily unavailable. the only catch: NVIDIA hasn’t listed GLM 5.3 yet, so for now you’ll be using GLM 5.2. once GLM 5.3 gets added to NVIDIA’s model catalog, you should be able to switch to it the same way.

🚨 GLM 5.3 just dropped, and here’s how to access the GLM family completely FREE. no subscription. no API credit cost & no rate limit. step 1️⃣: download the OpenCode app on your PC. it’s available on both Mac and Windows. 🔗: step 2️⃣: go to NVIDIA’s website, sign up with your email, verify your account, and generate a unified API key for free. NVIDIA offers free model endpoints for developers, and GLM 5.2 is one of the top models currently available there. 🔗: step 3️⃣: launch OpenCode, open the model selection menu, search for NVIDIA, and paste the API key you generated. also, there’s no rate limit, but traffic to the model through NVIDIA can get busy at times, so responses may occasionally be slower or temporarily unavailable. the only catch: NVIDIA hasn’t listed GLM 5.3 yet, so for now you’ll be using GLM 5.2. once GLM 5.3 gets added to NVIDIA’s model catalog, you should be able to switch to it the same way.

m0h

101,243 views • 2 months ago