01.AI on NVIDIA Ampere

Can I run Yi-1.5 34B on an RTX A6000?

Yes. Yi-1.5 34B at Q4_K_M uses 22.2 GB of the 44.3 GB available on a RTX A6000, and runs at about 23.3 tokens per second. You can push the context to 32K.

Runs comfortably

22.2 GB of 44.3 GB · 50%
044 GB
Weights 19.5 GB
KV cache 1.9 GB
Runtime overhead 0.9 GB

Yi-1.5 34B at Q4_K_M leaves 22.1 GB spare on a RTX A6000. There is room to raise the context length or move up a quantisation level.

Generation23.3tok/s
Prompt processing458tok/s
Max context32Ktokens
KV per 1K tokens0GB

Every quantisation of Yi-1.5 34B on a RTX A6000

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

QuantisationbpwWeights TotalVerdictMax ctxtok/s
FP16 / BF16 16.00 64.1 GB 66.8 GB Won't fit — 1.46
INT8 / W8A8 8.50 33.8 GB 36.6 GB Runs comfortably 32K 13.8
Q8_0 (GGUF) 8.50 33.8 GB 36.6 GB Runs comfortably 32K 13.8
FP8 (E4M3) 8.00 32.0 GB 34.8 GB Runs comfortably 32K 14.5
Q6_K 6.56 26.3 GB 29.0 GB Runs comfortably 32K 17.6
Q5_K_M 5.67 22.7 GB 25.5 GB Runs comfortably 32K 20.2
Q5_K_S 5.52 22.1 GB 24.9 GB Runs comfortably 32K 20.7
Q4_K_M 4.85 19.5 GB 22.2 GB Runs comfortably 32K 23.3
Q4_K_S 4.58 18.4 GB 21.2 GB Runs comfortably 32K 24.6
Q4_0 4.55 18.3 GB 21.1 GB Runs comfortably 32K 24.7
AWQ 4-bit 4.25 18.3 GB 21.0 GB Runs comfortably 32K 24.8
GPTQ 4-bit 4.25 18.3 GB 21.0 GB Runs comfortably 32K 24.8
MXFP4 4.25 18.3 GB 21.0 GB Runs comfortably 32K 24.8
IQ4_XS 4.25 17.2 GB 19.9 GB Runs comfortably 32K 26.3
Q3_K_M 3.91 15.8 GB 18.6 GB Runs comfortably 32K 28.4
IQ3_M 3.70 15.0 GB 17.8 GB Runs comfortably 32K 29.8
IQ3_XXS 3.06 12.5 GB 15.3 GB Runs comfortably 32K 35.3
Q2_K 2.63 10.8 GB 13.6 GB Runs comfortably 32K 40.2
IQ2_XXS 2.06 8.6 GB 11.4 GB Runs comfortably 32K 49.4
IQ1_M 1.75 7.4 GB 10.2 GB Runs comfortably 32K 56.4

Also worth checking

Yi-1.5 34B on every GPU
Full hardware tableEverything that runs on a RTX A6000
Full model tableWhy the KV cache is the part that bites
Guide