Skip to content

GLM 5.2 on Dual DGX Spark: VRAM, speed and fit

GLM 5.2 at 1.6-bit runs on the DGX Spark 128 GB with 864K of context at about 20 tokens a second.

Text generation · Chat · Ollama / LM Studio

Runs comfortably · on DGX Spark 128 GB · 1.6-bit (IQ1_S) · Ollama / LM Studio · Chat

GPU memory
244 GB
of 256 GB
Context
864K
auto
Speed
20 tok/s
2.0 when full
First reply
5.4 s
2K prompt

Where the memory goes

Model
145 GB
Context
82.3 GB
Overhead
1.2 GB
OS
15.4 GB
Free
12.4 GB
Memory
244 GB of 256 GB
Context
864K tokens · auto
Speed
20 tok/s · 2.0 when full
First reply
5.4 s for a 2K prompt
Running cost
$3.15/mo at home · $574/mo on a cloud 8× A40

Validation: Validated with conditions

What GLM 5.2 on Dual DGX Spark was checked against
CheckTargetResultStatus
Fits in GPU memory241 GB usable228 GB neededPass
Headroom5% free12.4 GB free (5%)Pass
Context for chat16K or more864KPass
Writing speed15 tok/s or more20 tok/sPass
First replyunder 3 s, 2K prompt5.4 sCheck

Targets: a reply you can read along with, a first reply the use case tolerates, and 5% of memory left free. Calculated from published specs, not benchmarked.

The setup

Model
GLM 5.2
Parameters
753B, 41.8B active
Precision
1.6-bit (IQ1_S)
Context
864K, auto
Chats at once
1
Use
Chat
Runs in
Ollama / LM Studio
Hardware
DGX Spark 128 GB
GPU can use
241 GB of 256 GB

Running cost

At home
$3.15/mo
144 W at the wall
Renting instead
$574/mo
8× A40 at $4.72/h

4 h a day, power at $0.18 per kWh, cloud at RunPod on-demand prices as of Nov 22, 2026.

How to run it

llama.cpp
llama-server -hf unsloth/GLM-5.2-GGUF
vLLM
vllm serve zai-org/GLM-5.2
MLX (Mac)
mlx_lm.chat --model mlx-community/GLM-5.2-DQ4plus-q8

GLM 5.2 on similar machines

All machines: GLM 5.2 VRAM requirements. Everything that runs here: what can the Dual DGX Spark run?

Other models that run on Dual DGX Spark

Questions

Can GLM 5.2 run on Dual DGX Spark?

GLM 5.2 at 1.6-bit runs on the DGX Spark 128 GB with 864K of context at about 20 tokens a second.

Report VR-1J38DB7, calculated Nov 22, 2026 from published specs, not benchmarked.