Zhipu on NVIDIA Ada

Can I run GLM-4 9B on an RTX 4060 Ti 16GB?

Yes. GLM-4 9B at Q4_K_M uses 6.5 GB of the 14.2 GB available on a RTX 4060 Ti 16GB, and runs at about 30.8 tokens per second. You can push the context to 128K.

Runs comfortably

6.5 GB of 14.2 GB · 46%
014 GB
Weights 5.4 GB
KV cache 0.3 GB
Runtime overhead 0.8 GB

GLM-4 9B at Q4_K_M leaves 7.7 GB spare on a RTX 4060 Ti 16GB. There is room to raise the context length or move up a quantisation level.

Generation30.8tok/s
Prompt processing983tok/s
Max context128Ktokens
KV per 1K tokens0GB

Every quantisation of GLM-4 9B on a RTX 4060 Ti 16GB

Highlighted row is the highest quality that still fits at 8K context.

QuantisationbpwWeights TotalVerdictMax ctxtok/s
FP16 / BF16 16.00 17.5 GB 18.6 GB Won't fit — 4.97
INT8 / W8A8 8.50 9.0 GB 10.2 GB Runs comfortably 113K 18.7
Q8_0 (GGUF) 8.50 9.0 GB 10.2 GB Runs comfortably 113K 18.7
FP8 (E4M3) 8.00 8.8 GB 9.9 GB Runs comfortably 119K 19.3
Q6_K 6.56 7.2 GB 8.3 GB Runs comfortably 128K 23.4
AWQ 4-bit 4.25 6.3 GB 7.5 GB Runs comfortably 128K 26.4
GPTQ 4-bit 4.25 6.3 GB 7.5 GB Runs comfortably 128K 26.4
MXFP4 4.25 6.3 GB 7.5 GB Runs comfortably 128K 26.4
Q5_K_M 5.67 6.2 GB 7.3 GB Runs comfortably 128K 27.0
Q5_K_S 5.52 6.0 GB 7.2 GB Runs comfortably 128K 27.7
Q4_K_M 4.85 5.4 GB 6.5 GB Runs comfortably 128K 30.8
Q4_K_S 4.58 5.1 GB 6.3 GB Runs comfortably 128K 32.3
Q4_0 4.55 5.1 GB 6.2 GB Runs comfortably 128K 32.5
IQ4_XS 4.25 4.8 GB 6.0 GB Runs comfortably 128K 34.3
Q3_K_M 3.91 4.5 GB 5.6 GB Runs comfortably 128K 36.7
IQ3_M 3.70 4.3 GB 5.4 GB Runs comfortably 128K 38.3
IQ3_XXS 3.06 3.7 GB 4.8 GB Runs comfortably 128K 44.3
Q2_K 2.63 3.3 GB 4.4 GB Runs comfortably 128K 49.5
IQ2_XXS 2.06 2.8 GB 3.9 GB Runs comfortably 128K 58.6
IQ1_M 1.75 2.5 GB 3.6 GB Runs comfortably 128K 65.1

Also worth checking

GLM-4 9B on every GPU
Full hardware tableEverything that runs on a RTX 4060 Ti 16GB
Full model tableWhy the KV cache is the part that bites
Guide