Only with CPU offload

Qwen3 32B at Q4_K_M needs 21.0 GB but only 14.4 GB is addressable, so about 36% of the layers would stream from system RAM at roughly 60 GB/s. Expect around 4.9 tokens per second — usable for batch work, painful for chat.

64% on GPU 8K context Q4_K_M · 18.4 GB Apache 2.0 Released Apr 2025

The classic 24 GB target, and still the strongest local translator under 70B. Qwen3.8 27B is smaller and better at everything else.

Hardware requirements What hardware do I need to run Qwen3 32B? →

At Q4_K_M with an 8K context: a 32 GB GPU (RTX 5090) or a 36 GB Mac (M4 Max). Minimum, recommended, high-quality and long-context tiers, with all 114 devices checked at every quantisation.

Fits instead: Qwen3 14B (10.2 GB) · Qwen3 8B (6.3 GB)

The VRAM budget

weights 18.4 GB
Weights 18.4 GB KV cache @ 8K 2.00 GB Runtime overhead 0.6 GB Over budget 6.6 GB past 14.4 GB

Quantisation ladder

QuantWeightsTotal @ 8KMax contextTok/sQualityFit
Q8_0 32.5 GB 35.1 GB — ~1.7 −0.1% ppl 20.7 GB over
Q6_K 25.0 GB 27.6 GB — ~2.6 −0.4% ppl 13.2 GB over
Q5_K_M 21.7 GB 24.3 GB — ~3.4 −0.8% ppl 9.9 GB over
Q4_K_M 18.4 GB 21.0 GB — ~4.9 −1.9% ppl 6.6 GB over
Q3_K_M 14.9 GB 17.5 GB — ~9.3 −5.4% ppl 3.1 GB over
Q2_K 12.8 GB 15.4 GB 4K ~20 −15% ppl 1.0 GB over

Quality is the published perplexity delta against f16 weights. Max context assumes an f16 KV cache; q8_0 roughly doubles it.

How to run it

terminal
$ llama-server \
    -hf Qwen/Qwen3-32B:Q4_K_M \
    -c 8192 -ngl 40

The engine underneath most of the others. Every knob is exposed. More on llama.cpp.

01Download is 18.4 GB. Keep it on an SSD — a first load off a spinning disk takes minutes.
02Close anything else holding VRAM. A browser with hardware acceleration can sit on 1–2 GB.
03If it falls back to CPU silently, drop the context first, then step down a quantisation.
See all models for this rig Compare with another model

Related models on a GeForce RTX 5070 Ti

ModelTotalTok/s est.
Qwen3 30B-A3B
30.5B MoE · same family
Q4_K_M · 18.5 GB ~26 CPU offload
Qwen3 Coder 30B-A3B
30.5B MoE · same family
Q4_K_M · 18.5 GB ~26 CPU offload
Qwen3 14B
14.8B · same family
Q4_K_M · 10.2 GB 65 Runs great
Qwen3 8B
8.19B · same family
Q4_K_M · 6.3 GB 118 Runs great
DeepSeek-R1-Distill-Qwen 32B
32.8B · similar size
Q4_K_M · 21.0 GB ~4.9 CPU offload
Qwen2.5-Coder 32B
32.8B · similar size
Q4_K_M · 21.0 GB ~4.9 CPU offload
LLM-jp 4 33B Thinking
33.2B · similar size
Q4_K_M · 21.3 GB ~4.7 CPU offload

Every model against the GeForce RTX 5070 Ti →

Qwen3 32B on other hardware

DeviceTotalTok/s est.
GeForce RTX 4060 Laptop
7.0 GB usable · 256 GB/s
21.0 GB ~2.4 CPU offload
GeForce RTX 3060 12 GB
10.6 GB usable · 360 GB/s
21.0 GB ~3.1 CPU offload
M4 · 16 GB
10.7 GB usable · 120 GB/s
21.0 GB — Won't fit
GeForce RTX 4090
22.4 GB usable · 1008 GB/s
21.0 GB 33 Tight fit
Radeon RX 7900 XTX
22.4 GB usable · 960 GB/s
21.0 GB 32 Tight fit
GeForce RTX 5090
30.4 GB usable · 1792 GB/s
21.0 GB 59 Runs great
M4 Max · 36 GB
27.0 GB usable · 410 GB/s
21.0 GB 12 Runs great
M3 Max · 64 GB
48.0 GB usable · 400 GB/s
21.0 GB 12 Runs great

All 114 devices, by tier →