The real cost to run GLM 5.2 NVFP4

lukealonso · 432B parameters · 1M context · commercial OK

GLM 5.2 NVFP4 — 432B params (lukealonso). Auto-indexed from the Hugging Face Hub (40,291 downloads). Parameter count is exact; download size and quantizations are estimates.

What it costs to run GLM 5.2 NVFP4 — $0 markup

On your own machine — $0. Runs free locally if you have about 322 GB of VRAM at Q4_K_M (point Spanvero at LM Studio, Ollama, or llama.cpp).
On your own rented GPU — from $3.92/hr. 8× NVIDIA RTX A6000 48GB at the direct RunPod price (≈ $5.51 a session), $0 markup.
Via your own API key — $3.56/1M tokens (blended input + output). No GPU to manage (rough estimate for this size).

Key facts

Parameters	432B
Context window	1M tokens
Recommended quant	Q4_K_M
VRAM to run	~322 GB (at Q4_K_M, 16.4K context)
Download size	~259 GB
License	Commercial use OK

Open the free Spanvero advisor → for the live, interactive math for your exact workload and hardware.

Related models

Llama 3.1 405B — 405.9B, meta-llama
Llama 3.1 405B Instruct — 405B, Meta
Llama 4 Maverick (17B-128E) — 402B, Meta
DeepSeek R1 0528 NVFP4 v2 — 393.6B, nvidia
Qwen3 Coder 480B A35B Instruct — 480.2B, Qwen

Browse: All models · Compare