DeepSeek on NVIDIA Ampere

Can I run DeepSeek-R1-Distill-Qwen 32B on an RTX 3080 10GB?

Not at Q4_K_M — it needs 21.5 GB against 8.6 GB available. You would need 3 of these cards.

Won't fit

21.5 GB of 8.6 GB · 249%
022 GB
Weights 18.6 GB
KV cache 2.0 GB
Runtime overhead 0.8 GB
Over the limit 12.9 GB

Short by 12.9 GB. You can run it with 19 of 64 layers on the RTX 3080 10GB and the rest in system RAM, at roughly 2.79 tok/s — usable for batch work, painful for chat. A smaller quantisation or a shorter context is usually the better trade.

Generation2.79tok/s
Prompt processing385tok/s
Max context0tokens
KV per 1K tokens0GB

Every quantisation of DeepSeek-R1-Distill-Qwen 32B on a RTX 3080 10GB

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

QuantisationbpwWeights TotalVerdictMax ctxtok/s
FP16 / BF16 16.00 61.0 GB 63.9 GB Won't fit — 0.70
INT8 / W8A8 8.50 32.1 GB 34.9 GB Won't fit — 1.43
Q8_0 (GGUF) 8.50 32.1 GB 34.9 GB Won't fit — 1.43
FP8 (E4M3) 8.00 30.5 GB 33.3 GB Won't fit — 1.53
Q6_K 6.56 25.0 GB 27.9 GB Won't fit — 1.92
Q5_K_M 5.67 21.6 GB 24.5 GB Won't fit — 2.33
Q5_K_S 5.52 21.1 GB 23.9 GB Won't fit — 2.39
Q4_K_M 4.85 18.6 GB 21.5 GB Won't fit — 2.79
AWQ 4-bit 4.25 18.3 GB 21.2 GB Won't fit — 2.89
GPTQ 4-bit 4.25 18.3 GB 21.2 GB Won't fit — 2.89
MXFP4 4.25 18.3 GB 21.2 GB Won't fit — 2.89
Q4_K_S 4.58 17.6 GB 20.5 GB Won't fit — 3.00
Q4_0 4.55 17.5 GB 20.4 GB Won't fit — 3.08
IQ4_XS 4.25 16.4 GB 19.3 GB Won't fit — 3.34
Q3_K_M 3.91 15.2 GB 18.0 GB Won't fit — 3.75
IQ3_M 3.70 14.4 GB 17.3 GB Won't fit — 4.02
IQ3_XXS 3.06 12.1 GB 15.0 GB Won't fit — 5.33
Q2_K 2.63 10.6 GB 13.4 GB Won't fit — 6.72
IQ2_XXS 2.06 8.5 GB 11.3 GB Won't fit — 10.9
IQ1_M 1.75 7.4 GB 10.2 GB Won't fit 2K 16.7

Also worth checking

DeepSeek-R1-Distill-Qwen 32B on every GPU
Full hardware tableEverything that runs on a RTX 3080 10GB
Full model tableWhy the KV cache is the part that bites
Guide