01.AI on NVIDIA Ada

Can I run Yi-1.5 34B on an RTX 4080 Super?

Not at Q4_K_M — it needs 22.2 GB against 14.2 GB available. You would need 2 of these cards.

Won't fit

22.2 GB of 14.2 GB · 156%
023 GB
Weights 19.5 GB
KV cache 1.9 GB
Runtime overhead 0.9 GB
Over the limit 8.0 GB

Short by 8.0 GB. You can run it with 35 of 60 layers on the RTX 4080 Super and the rest in system RAM, at roughly 4.00 tok/s — usable for batch work, painful for chat. A smaller quantisation or a shorter context is usually the better trade.

Generation4.00tok/s
Prompt processing635tok/s
Max context0tokens
KV per 1K tokens0GB

Every quantisation of Yi-1.5 34B on a RTX 4080 Super

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 — 0.69
INT8 / W8A8 8.50 33.8 GB 36.6 GB Won't fit — 1.58
Q8_0 (GGUF) 8.50 33.8 GB 36.6 GB Won't fit — 1.58
FP8 (E4M3) 8.00 32.0 GB 34.8 GB Won't fit — 1.71
Q6_K 6.56 26.3 GB 29.0 GB Won't fit — 2.33
Q5_K_M 5.67 22.7 GB 25.5 GB Won't fit — 2.98
Q5_K_S 5.52 22.1 GB 24.9 GB Won't fit — 3.14
Q4_K_M 4.85 19.5 GB 22.2 GB Won't fit — 4.00
Q4_K_S 4.58 18.4 GB 21.2 GB Won't fit — 4.51
Q4_0 4.55 18.3 GB 21.1 GB Won't fit — 4.54
AWQ 4-bit 4.25 18.3 GB 21.0 GB Won't fit — 4.55
GPTQ 4-bit 4.25 18.3 GB 21.0 GB Won't fit — 4.55
MXFP4 4.25 18.3 GB 21.0 GB Won't fit — 4.55
IQ4_XS 4.25 17.2 GB 19.9 GB Won't fit — 5.40
Q3_K_M 3.91 15.8 GB 18.6 GB Won't fit — 6.59
IQ3_M 3.70 15.0 GB 17.8 GB Won't fit — 7.60
IQ3_XXS 3.06 12.5 GB 15.3 GB Won't fit 4K 17.4
Q2_K 2.63 10.8 GB 13.6 GB Fits, but tight 11K 37.0
IQ2_XXS 2.06 8.6 GB 11.4 GB Runs comfortably 20K 45.4
IQ1_M 1.75 7.4 GB 10.2 GB Runs comfortably 25K 51.9

Also worth checking

Yi-1.5 34B on every GPU
Full hardware tableEverything that runs on a RTX 4080 Super
Full model tableWhy the KV cache is the part that bites
Guide