How much VRAM does GLM-5.2 753B (MoE) need at Q4_K_M? Zhipu 1M MLA frontier
By TechCompare · Updated · How we check
GLM-5.2 753B at Q4_K_M is a 648 GB resident deployment for 1M context. The 422 GB of weights are the biggest piece, and the 78-layer MLA cache adds 168 GB on top.
Calculator
Estimated VRAM required
648 GB
753B params at Q4_K_M, 1,048,576 token context, batch 1, inference.
Estimate accuracy: Weights within ~2%. KV cache within ~5% for standard GQA models, ~10% for MLA or hybrid state estimates. Real VRAM may vary with framework (vLLM vs llama.cpp vs Transformers), Flash Attention, and driver overhead.
Hardware that fits
Apple entries show installed unified memory separately from a model budget based on an 8 GB reserve. That budget is a planning heuristic, not a fixed hardware limit.
No single GPU in our catalog has enough memory. Multi-GPU or CPU offload required.
GLM-5.2 753B at Q4_K_M with native 1M context needs about 648 GB of VRAM with all experts resident. Zhipu's successor to GLM-5.1 activates 40B parameters per token from a 753B total pool and rebuilds attention as 78 MLA layers interleaved with DSA IndexShare every 4 layers. Those 40B active params don't shrink what has to stay loaded, because different tokens pick different experts.
How this is calculated
GLM-5.2 ships at 753B total parameters on HuggingFace config, close to GLM-5.1's 754B pool. The attention is rebuilt: 78 MLA layers compressed to a 512-dim latent with 1 effective KV head, plus DSA IndexShare every 4 layers for sequence sparsity.
At Q4_K_M the full parameter pool is 422 GB of weights. The MLA key-value cache at the full 1M context window uses 168 GB. Activation and software overhead adds about 59 GB. The cache depends on context length, not on which experts a token picks.
Verdict
Use the official Zhipu API unless data residency is the whole reason for self-hosting. Resident mode needs 9 cards of 80 GB or 5 H200 cards of 141 GB, both with NVLink. The cache is the part you control: at 256K it falls to roughly 42 GB.
More GLM scenarios
Related guides
Frequently asked questions
How much of GLM-5.2's footprint is KV cache at 1M context?
How is GLM-5.2 different from GLM-5.1?
Is a 1M-token context window worth the VRAM cost?
Related tools
RAM Latency Calculator
Convert DDR3/DDR4/DDR5 timings (CL, tRCD, tRP, tRAS) into true latency in nanoseconds.
Use toolPower Cost Estimator
Estimate annual electricity costs for your PC, Server, or TV.
Use toolData Transfer Calculator
Estimate transfer times for files over USB, WiFi, Ethernet, and more.
Use toolMemory and Storage Throughput Visualizer
Visualize the large speed gap between CPU cache, RAM, and storage.
Use tool