RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can GLM 5.2 run on the L40S?

No. Even at 2-bit with 8K context, GLM 5.2 needs about 265 GB, more than the 45.6 GB usable on the L40S.

Memory by quant and context

Quant4K ctx8K ctx32K ctx128K ctxMax tok/s (est., 8K)
2-bit (Q2_K)264 GB No265 GB No267 GB No277 GB No—
3-bit (Q3_K_M) •320 GB No321 GB No323 GB No333 GB No—
4-bit (Q4_K_M) •435 GB No435 GB No438 GB No448 GB No—
5-bit (Q5_K_M) •523 GB No524 GB No526 GB No536 GB No—
6-bit (Q6_K) •584 GB No584 GB No587 GB No597 GB No—
8-bit (Q8_0) •747 GB No748 GB No750 GB No760 GB No—
16-bit (F16/BF16) •1406 GB No1406 GB No1408 GB No1419 GB No—

• real GGUF file size from unsloth/GLM-5.2-GGUF. Speed is a bandwidth ceiling (864 GB/s), not a benchmark. Methodology.

Hardware that does fit it

Mac M3 Ultra 512GB

More

GLM 5.2 on every GPU and Mac · Every model that runs on the L40S · Other models that fit 48 GB VRAM · Try other settings in the calculator

Data: Hugging Face Hub (zai-org/GLM-5.2), NVIDIA specs. Data as of 2026-11-17.