Loading video...

Video Failed to Load

Go Home

GLM-5.2 Nvidia NVFP4 No pruning, no quantising, exact same model card. 110 tok/s single stream. My friend is a genius, holy fudge. Here's the repo, he got full DS4-Flash on a 5090 + DDR5 at 38.5 tok/s This is a breakthrough.

0xSero's profile picture

0xSero

64,603 subscribers

157,562 views • 2 months ago •via X (Twitter)

Science & Technology
LIVE

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

GLM-5.3-Flash running at up to 300 tok/s single stream, and 1000 tok/s with 8 concurrent sessions. On average it's working at 150 tok/s per stream. Phenomenal model, really something special. On 2x Sparks we get 60 tok/s and 4x 100 tok/s 300B params is the sweet spot.
1:10

GLM-5.3-Flash running at up to 300 tok/s single stream, and 1000 tok/s with 8 concurrent sessions. On average it's working at 150 tok/s per stream. Phenomenal model, really something special. On 2x Sparks we get 60 tok/s and 4x 100 tok/s 300B params is the sweet spot.

0xSero

30,601 views • 22 days ago

The RTX 3090 is a 5-year-old GPU and it still runs a 27B model at 20 tok/s I tested Qwen3.5:27b across 3 generations of NVIDIA: 5090 → ~60 tok/s 4090 → ~40 tok/s 3090 → ~20 tok/s Perfectly linear scaling. Double the generation, double the speed.
0:57

The RTX 3090 is a 5-year-old GPU and it still runs a 27B model at 20 tok/s I tested Qwen3.5:27b across 3 generations of NVIDIA: 5090 → ~60 tok/s 4090 → ~40 tok/s 3090 → ~20 tok/s Perfectly linear scaling. Double the generation, double the speed.

stevibe

142,493 views • 6 months ago

GLM-5.1-478B-NVFP4 Running on: - 4x RTX Pro 6000 - Sglang - 370,000 max tokens (1.75x full context) - p10 27.7 | p90 45.6 tok/s decode (gen) - 1340 tok/s prefill I could get 2x decode if I limit to 64k context (100 tok/s) In this video it operates Figma (:
1:00

GLM-5.1-478B-NVFP4 Running on: - 4x RTX Pro 6000 - Sglang - 370,000 max tokens (1.75x full context) - p10 27.7 | p90 45.6 tok/s decode (gen) - 1340 tok/s prefill I could get 2x decode if I limit to 64k context (100 tok/s) In this video it operates Figma (:

0xSero

74,772 views • 5 months ago

Nemotron-3-Ultra running on 4x 6000s edits my latest demo video.. - 75 tok/s decode - 8x concurrency - 256k context - 899 tok/s prefill - 20k tok/s prefill cache - NVFP4 Setting it up to be my Hermes driver. It's good enough at most things and doesn't talk like a moron.
1:26

Nemotron-3-Ultra running on 4x 6000s edits my latest demo video.. - 75 tok/s decode - 8x concurrency - 256k context - 899 tok/s prefill - 20k tok/s prefill cache - NVFP4 Setting it up to be my Hermes driver. It's good enough at most things and doesn't talk like a moron.

0xSero

15,674 views • 3 months ago

Qwen3.5-35B with only 3B active parameters This MoE model runs FASTER than most 7B dense models. Tested on 3 generations of NVIDIA: - 5090: 137 tok/s - 4090: 112 tok/s - 3090: 78 tok/s The surprise? The 4090 <> 5090 gap is only 22%. With a 3B active MoE, even old GPUs fly.
0:30

Qwen3.5-35B with only 3B active parameters This MoE model runs FASTER than most 7B dense models. Tested on 3 generations of NVIDIA: - 5090: 137 tok/s - 4090: 112 tok/s - 3090: 78 tok/s The surprise? The 4090 <> 5090 gap is only 22%. With a 3B active MoE, even old GPUs fly.

stevibe

69,532 views • 6 months ago

Finally!! Achieved 385 tok/s with GLM 5.2 NVFP4 on 8x RTX PRO 6000. It was buggy and incredibly slow at first. It took me over a week using agents to get this far.
0:22

Finally!! Achieved 385 tok/s with GLM 5.2 NVFP4 on 8x RTX PRO 6000. It was buggy and incredibly slow at first. It took me over a week using agents to get this far.

Hikari∣LocalLLM⚡

103,398 views • 1 month ago

Minimax-M3 running on 4x RTX Pro 6000s - 800k context - 4x concurrency at 250k - 70-120 tok/s - 2000 tok/s prefill no cache - 376gb vram - mxfp4 It's working on improving the audio on one of my videos, it's actually doing a good job in researching solutions. Good model
0:33

Minimax-M3 running on 4x RTX Pro 6000s - 800k context - 4x concurrency at 250k - 70-120 tok/s - 2000 tok/s prefill no cache - 376gb vram - mxfp4 It's working on improving the audio on one of my videos, it's actually doing a good job in researching solutions. Good model

0xSero

24,634 views • 3 months ago

GLM 4.7 Flash is supported in mlx-lm 0.30.3 (h/t Ivan Fioravanti ᯅ) The 4-bit runs fast (43 tok/s generation, ~800 tok/s prefill) on a base M5 32GB laptop.
0:37

GLM 4.7 Flash is supported in mlx-lm 0.30.3 (h/t Ivan Fioravanti ᯅ) The 4-bit runs fast (43 tok/s generation, ~800 tok/s prefill) on a base M5 32GB laptop.

Awni Hannun

141,194 views • 8 months ago

125B on a single RTX 4090 at 25 tok/s is actually insane Qwen3.8-Flash-Next + MTP is hitting: 80K context 25.35 tok/s decode 471 tok/s prefill On one consumer GPU. Yes, it’s an MoE, so you’re not running all 125B parameters for every token. But still, getting a model this large to run this fast on a 4090 is wild. MTP is doing a lot of the heavy lifting here too. Instead of generating one token at a time, it can speculate on multiple tokens and verify them together. And 25 tok/s is the important number.
0:20

125B on a single RTX 4090 at 25 tok/s is actually insane Qwen3.8-Flash-Next + MTP is hitting: 80K context 25.35 tok/s decode 471 tok/s prefill On one consumer GPU. Yes, it’s an MoE, so you’re not running all 125B parameters for every token. But still, getting a model this large to run this fast on a 4090 is wild. MTP is doing a lot of the heavy lifting here too. Instead of generating one token at a time, it can speculate on multiple tokens and verify them together. And 25 tok/s is the important number.

FHILY👑

39,811 views • 16 days ago

Qwen3.8-Flash-Next now runs on a single RTX 5090 at 68.3 tok/s — no extreme quantization, no speculative decoding. · Production-level checkpoint: GB300-validated NVFP4 checkpoint from RadixArk · 63GB host RAM — less than what 1-bit quants of this model need · The 51GB n-gram table? It lives on your NVMe, at ~0.5% throughput cost In the video: one prompt → a playable minecraft-style world, then FreeToken on my 5090 desktop fixing a real FreeToken bug in 10.5 minutes. Meet FreeToken 🧵
0:58

Qwen3.8-Flash-Next now runs on a single RTX 5090 at 68.3 tok/s — no extreme quantization, no speculative decoding. · Production-level checkpoint: GB300-validated NVFP4 checkpoint from RadixArk · 63GB host RAM — less than what 1-bit quants of this model need · The 51GB n-gram table? It lives on your NVMe, at ~0.5% throughput cost In the video: one prompt → a playable minecraft-style world, then FreeToken on my 5090 desktop fixing a real FreeToken bug in 10.5 minutes. Meet FreeToken 🧵

Shuo Yang

62,820 views • 15 days ago

16 people chatting with Qwen3.6-35B at once ONE DGX Spark. This is a real capture, not a mockup: every token you see replays at its actual measured arrival time. Peaks ~440 tok/s total. A single user gets 105 tok/s. NVFP4 + MTP-3 on vLLM.
0:34

16 people chatting with Qwen3.6-35B at once ONE DGX Spark. This is a real capture, not a mockup: every token you see replays at its actual measured arrival time. Peaks ~440 tok/s total. A single user gets 105 tok/s. NVFP4 + MTP-3 on vLLM.

Wësche

26,094 views • 2 months ago

Everyone's comparing the DGX Spark to a 5090 and calling it slow. I think that's the wrong comparison. I ran Qwen3.6 35B-A3B FP8 with the full 262K context window enabled (~96GB RAM) — something gaming GPUs can't really do. Results: 🟢No context: 51.3 tok/s, TTFT 110ms 🟣200K prefill: 34.6 tok/s, TTFT 85s (~2,341 tok/s prefill) Prefill is way faster than a Mac. And 35 tok/s deep into 200K context, on a model this strong, is genuinely usable. The Spark plays a different game.
1:04

Everyone's comparing the DGX Spark to a 5090 and calling it slow. I think that's the wrong comparison. I ran Qwen3.6 35B-A3B FP8 with the full 262K context window enabled (~96GB RAM) — something gaming GPUs can't really do. Results: 🟢No context: 51.3 tok/s, TTFT 110ms 🟣200K prefill: 34.6 tok/s, TTFT 85s (~2,341 tok/s prefill) Prefill is way faster than a Mac. And 35 tok/s deep into 200K context, on a model this strong, is genuinely usable. The Spark plays a different game.

stevibe

34,141 views • 4 months ago

Kimi-K2.6 running locally 110+ tok/s
0:42

Kimi-K2.6 running locally 110+ tok/s

0xSero

24,640 views • 3 months ago

Qwen3.8-Flash-Next just got a massive speed boost on my old hardware. This guy started with roughly 11 tok/s. After tuning the inference stack, I’m now getting around 24–28 tok/s. That’s roughly a 2× speedup without changing the GPU. The setup: → RTX 3090 24GB → 64GB DDR4 system RAM → Qwen3.8-Flash-Next → Cafe-llama.cpp The interesting part is that Cafe-llama.cpp isn’t just another llama.cpp fork. The optimizations are making a real difference for this particular architecture, especially when you’re dealing with a model that relies heavily on system memory and offloading. Going from 11 tok/s to nearly 30 tok/s completely changes how practical the model feels for local use. And now I’m curious what happens with DDR5. If memory bandwidth becomes another bottleneck, a faster RAM subsystem could squeeze even more performance out of the same 3090. This is why local AI optimization is getting so interesting. Sometimes you don’t need a new GPU. You just need a better inference stack. 11 tok/s → 24–28 tok/s.
0:50

Qwen3.8-Flash-Next just got a massive speed boost on my old hardware. This guy started with roughly 11 tok/s. After tuning the inference stack, I’m now getting around 24–28 tok/s. That’s roughly a 2× speedup without changing the GPU. The setup: → RTX 3090 24GB → 64GB DDR4 system RAM → Qwen3.8-Flash-Next → Cafe-llama.cpp The interesting part is that Cafe-llama.cpp isn’t just another llama.cpp fork. The optimizations are making a real difference for this particular architecture, especially when you’re dealing with a model that relies heavily on system memory and offloading. Going from 11 tok/s to nearly 30 tok/s completely changes how practical the model feels for local use. And now I’m curious what happens with DDR5. If memory bandwidth becomes another bottleneck, a faster RAM subsystem could squeeze even more performance out of the same 3090. This is why local AI optimization is getting so interesting. Sometimes you don’t need a new GPU. You just need a better inference stack. 11 tok/s → 24–28 tok/s.

FHILY👑

23,243 views • 19 days ago

GLM-5.3 weights from Z.ai are live, with SGLang powering day-0 serving support! GLM-5.3 inherits every optimization and feature we battle-tested for GLM-5.2 over the past months. On real-world multi-turn agentic workloads, we measured 537.6 tok/s/user on NVFP4 and 413 tok/s/user on FP8, at BS=1 with TP8 on 8x B300. It's fast, efficient, and production-ready today on NVIDIA AI Blackwell and Hopper, and AI at AMD MI300X/325X/355X. We believe this is a big step forward for GLM-Series in agentic tasks, with a path toward mythos-class cyber capability. Can't wait to see what people build with GLM-5.3 and SGLang🚀 Cookbook👇
0:20

GLM-5.3 weights from Z.ai are live, with SGLang powering day-0 serving support! GLM-5.3 inherits every optimization and feature we battle-tested for GLM-5.2 over the past months. On real-world multi-turn agentic workloads, we measured 537.6 tok/s/user on NVFP4 and 413 tok/s/user on FP8, at BS=1 with TP8 on 8x B300. It's fast, efficient, and production-ready today on NVIDIA AI Blackwell and Hopper, and AI at AMD MI300X/325X/355X. We believe this is a big step forward for GLM-Series in agentic tasks, with a path toward mythos-class cyber capability. Can't wait to see what people build with GLM-5.3 and SGLang🚀 Cookbook👇

SGLang

99,165 views • 21 days ago

"Why are you benchmarking DGX Spark? It's a training box." Yeah. Low bandwidth, but 128GB of unified memory is just sitting there. Plenty of room to optimize. DGX Spark + Qwen3.6 27B. Four backend/quant combos: 🔴 llama.cpp + UD_Q4_K_XL > 11.0 tok/s (baseline), TTFT 297ms 🟢 llama.cpp + DFlash > 20.4 tok/s (peaks at 97 tok/s), TTFT 320ms 🟡 vLLM FP8 + MTP > 13.1 tok/s, TTFT 540ms 🟣 vLLM NVFP4 + MTP > 24.2 tok/s, TTFT 376ms NVFP4+MTP is the winner for me, rock stable around 24 tok/s, no wild swings. DFlash is the wildcard: massive peaks, but fluctuates a lot. FP8+MTP barely beats baseline, and it's FP8. Love my Spark.
3:01

"Why are you benchmarking DGX Spark? It's a training box." Yeah. Low bandwidth, but 128GB of unified memory is just sitting there. Plenty of room to optimize. DGX Spark + Qwen3.6 27B. Four backend/quant combos: 🔴 llama.cpp + UD_Q4_K_XL > 11.0 tok/s (baseline), TTFT 297ms 🟢 llama.cpp + DFlash > 20.4 tok/s (peaks at 97 tok/s), TTFT 320ms 🟡 vLLM FP8 + MTP > 13.1 tok/s, TTFT 540ms 🟣 vLLM NVFP4 + MTP > 24.2 tok/s, TTFT 376ms NVFP4+MTP is the winner for me, rock stable around 24 tok/s, no wild swings. DFlash is the wildcard: massive peaks, but fluctuates a lot. FP8+MTP barely beats baseline, and it's FP8. Love my Spark.

stevibe

40,223 views • 4 months ago

I am blown away by MiMo-V2.5 It's so steerable, I didn't like MiMo-V2 This is self hosted btw, 125 tok/s across 2 sessions 4k prefill tok/s It's updating my website + editing a longer video I will be posting in a few minutes.
0:25

I am blown away by MiMo-V2.5 It's so steerable, I didn't like MiMo-V2 This is self hosted btw, 125 tok/s across 2 sessions 4k prefill tok/s It's updating my website + editing a longer video I will be posting in a few minutes.

0xSero

34,494 views • 4 months ago

1-bit GLM-5.2 GGUF vs. Claude 4.8 Opus vs. GPT-5.5 We gave 3 models the same prompt and compared one-shot outputs. The 1-bit GLM-5.2 GGUF ran locally on a Mac Studio M3 Ultra with 256GB RAM at ~21.6 tok/s. Which output do you like best? GGUF:
0:21

1-bit GLM-5.2 GGUF vs. Claude 4.8 Opus vs. GPT-5.5 We gave 3 models the same prompt and compared one-shot outputs. The 1-bit GLM-5.2 GGUF ran locally on a Mac Studio M3 Ultra with 256GB RAM at ~21.6 tok/s. Which output do you like best? GGUF:

Unsloth AI

1,543,874 views • 2 months ago

Qwen3.6 35B-A3B dropped yesterday, so I ran it on 4 GPUs to see how it performs: 🟣 RTX 3090 — 49.78 tok/s, TTFT 852ms 🟡 RTX 4090 — 118.93 tok/s, TTFT 686ms 🟢 RTX 5090 — 160.37 tok/s, TTFT 409ms 🔵 DGX Spark — 59.98 tok/s, TTFT 228ms I went with ollama as the backend because honestly, it's the easiest way for most people to get started. One command, model pulled, done. I used Q4_K_M (24GB) across all four cards. The reason is the 3090 and 4090 don't support NVFP4 (only the 5090 and DGX Spark could use it). Keeping the same quant everywhere felt like the fairest way to compare. And yes, you can absolutely squeeze more performance out of every card with vLLM, SGLang, or TensorRT-LLM. But that's not what this test is about. This is just the out-of-the-box experience for folks who own a GPU and want to try the new model tonight.
1:24

Qwen3.6 35B-A3B dropped yesterday, so I ran it on 4 GPUs to see how it performs: 🟣 RTX 3090 — 49.78 tok/s, TTFT 852ms 🟡 RTX 4090 — 118.93 tok/s, TTFT 686ms 🟢 RTX 5090 — 160.37 tok/s, TTFT 409ms 🔵 DGX Spark — 59.98 tok/s, TTFT 228ms I went with ollama as the backend because honestly, it's the easiest way for most people to get started. One command, model pulled, done. I used Q4_K_M (24GB) across all four cards. The reason is the 3090 and 4090 don't support NVFP4 (only the 5090 and DGX Spark could use it). Keeping the same quant everywhere felt like the fairest way to compare. And yes, you can absolutely squeeze more performance out of every card with vLLM, SGLang, or TensorRT-LLM. But that's not what this test is about. This is just the out-of-the-box experience for folks who own a GPU and want to try the new model tonight.

stevibe

403,049 views • 5 months ago

GLM-5.3-Flash (320B MoE) serving on Kaggle's free TPU: ~64 tok/s, 262k context. No stack could run this model on a TPU, so I wrote a JAX engine for it: 3-bit experts (from Unsloth AI) in HBM; Pallas kernels decode them inside the matmul. Z.ai Kaggle
0:43

GLM-5.3-Flash (320B MoE) serving on Kaggle's free TPU: ~64 tok/s, 262k context. No stack could run this model on a TPU, so I wrote a JAX engine for it: 3-bit experts (from Unsloth AI) in HBM; Pallas kernels decode them inside the matmul. Z.ai Kaggle

Abdur Rahim

104,939 views • 3 days ago