01.AI on NVIDIA Blackwell

Can I run Yi-1.5 34B on an RTX 5090?

Yes. Yi-1.5 34B at Q4_K_M uses 22.2 GB of the 29.3 GB available on a RTX 5090, and runs at about 56.8 tokens per second. You can push the context to 32K.

Runs comfortably

22.2 GB of 29.3 GB · 76%
029 GB
Weights 19.5 GB
KV cache 1.9 GB
Runtime overhead 0.9 GB

Yi-1.5 34B at Q4_K_M leaves 7.0 GB spare on a RTX 5090. There is room to raise the context length or move up a quantisation level.

Generation56.8tok/s
Prompt processing1276tok/s
Max context32Ktokens
KV per 1K tokens0GB

Every quantisation of Yi-1.5 34B on a RTX 5090

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 — 1.06
INT8 / W8A8 8.50 33.8 GB 36.6 GB Won't fit — 4.95
Q8_0 (GGUF) 8.50 33.8 GB 36.6 GB Won't fit — 4.95
FP8 (E4M3) 8.00 32.0 GB 34.8 GB Won't fit — 6.01
Q6_K 6.56 26.3 GB 29.0 GB Fits, but tight 9K 42.9
Q5_K_M 5.67 22.7 GB 25.5 GB Runs comfortably 24K 49.2
Q5_K_S 5.52 22.1 GB 24.9 GB Runs comfortably 27K 50.5
Q4_K_M 4.85 19.5 GB 22.2 GB Runs comfortably 32K 56.8
Q4_K_S 4.58 18.4 GB 21.2 GB Runs comfortably 32K 59.8
Q4_0 4.55 18.3 GB 21.1 GB Runs comfortably 32K 60.2
AWQ 4-bit 4.25 18.3 GB 21.0 GB Runs comfortably 32K 60.3
GPTQ 4-bit 4.25 18.3 GB 21.0 GB Runs comfortably 32K 60.3
MXFP4 4.25 18.3 GB 21.0 GB Runs comfortably 32K 60.3
IQ4_XS 4.25 17.2 GB 19.9 GB Runs comfortably 32K 64.0
Q3_K_M 3.91 15.8 GB 18.6 GB Runs comfortably 32K 68.9
IQ3_M 3.70 15.0 GB 17.8 GB Runs comfortably 32K 72.3
IQ3_XXS 3.06 12.5 GB 15.3 GB Runs comfortably 32K 85.3
Q2_K 2.63 10.8 GB 13.6 GB Runs comfortably 32K 97.1
IQ2_XXS 2.06 8.6 GB 11.4 GB Runs comfortably 32K 119
IQ1_M 1.75 7.4 GB 10.2 GB Runs comfortably 32K 135

Also worth checking

Yi-1.5 34B on every GPU
Full hardware tableEverything that runs on a RTX 5090
Full model tableWhy the KV cache is the part that bites
Guide