Can A100 80GB SXM4 run GLM-4.5?

NVIDIA · 80 GB HBM2e · 2039 GB/s bandwidth · Z.ai · 358.3B (active 32B) · MoE · model ctx up to 128K

No — GLM-4.5 at Q4 needs ≈ 201.27 GB even at 4K context, beyond the A100 80GB SXM4's 80 GB HBM2e. It would only run with heavy CPU/disk offload. Decode at Q4/8K: ≈ 17.7 tok/s (estimated, batch 1). FP16 needs ≈ 724.0 GB — does not fit on this card. Estimates come from memory-bandwidth math; rows tagged measured override estimates. Relative ranking is more reliable than absolute tok/s. Methodology.

Fit & speed by quant and context

Quant	Context	VRAM needed	Fits?	tok/s (decode)
Q4	4K	201.27 GB	No	18.3 est.
Q4	8K default	204.47 GB	No	17.7 est.
Q4	32K	223.67 GB	No	14.8 est.
FP16	4K	720.8 GB	No	—
FP16	8K default	724.0 GB	No	—
FP16	32K	743.2 GB	No	—

Fits = weights + KV(ctx) + 1 GB overhead ≤ 95% of VRAM. Measured anchors are context-agnostic; the fit verdict is recomputed per context. A missing tok/s means the model is far beyond this card (offload-only territory).

VRAM breakdown at 8K context

Quant	Weights	KV cache	Overhead	Total needed	A100 80GB SXM4 VRAM
Q4	197.07 GB	6.4 GB	1 GB	204.47 GB	80 GB
FP16	716.6 GB	6.4 GB	1 GB	724.0 GB	80 GB

Other GPUs that run GLM-4.5

No catalog GPU fully fits GLM-4.5 at Q4/8K — see the model page for offload estimates.

All GPUs for GLM-4.5 →

Other models for the A100 80GB SXM4

• Qwen3-0.6B — ≈1621.9 tok/s Q4 est.
• gemma-3-1b-it — ≈1297.5 tok/s Q4 est.
• Llama-3.2-1B-Instruct — ≈1081.3 tok/s Q4 est.
• Qwen3-1.7B — ≈648.8 tok/s Q4 est.
• Qwen3-30B-A3B — ≈432.5 tok/s Q4 est.
• Qwen3-Coder-30B-A3B-Instruct — ≈432.5 tok/s Q4 est.
• GLM-4.7-Flash — ≈432.5 tok/s Q4 est.
• Qwen3-30B-A3B-Instruct-2507 — ≈432.5 tok/s Q4 est.

All models on the A100 80GB SXM4 →