Yes — with 1.3 GB to spare

Qwen3 4B at Q4_K_M fits your GeForce GTX 1660 Super entirely on the GPU at 8K context, at an estimated 90 tokens per second. Past 17K the KV cache pushes it over — quantise the cache to q8_0, or step down a quantisation, to go longer.

Fully on GPU 8K context Q4_K_M · 2.3 GB Apache 2.0 Released Apr 2025

The 2025 sweet spot for 8 GB cards with reasoning traces. Qwen3.5 4B adds vision and 8× the context.

Hardware requirements What hardware do I need to run Qwen3 4B? →

At Q4_K_M with an 8K context: a 6 GB GPU (RTX 3060 Laptop) or a 8 GB Mac (M3). Minimum, recommended, high-quality and long-context tiers, with all 114 devices checked at every quantisation.

The VRAM budget

weights 2.3 GB
Weights 2.3 GB KV cache @ 8K 1.13 GB Runtime overhead 0.6 GB Free 1.3 GB of 5.3 GB

Quantisation ladder

QuantWeightsTotal @ 8KMax contextTok/sQualityFit
F16 7.5 GB 9.2 GB — ~8.0 Reference 3.9 GB over
Q8_0 4.0 GB 5.7 GB 5K ~35 −0.1% ppl 0.4 GB over
Q6_K 3.1 GB 4.8 GB 11K 66 −0.4% ppl Fits
Q5_K_M 2.7 GB 4.4 GB 14K 77 −0.8% ppl Fits
Q4_K_M 2.3 GB 4.0 GB 17K 90 −1.9% ppl Recommended
Q3_K_M 1.8 GB 3.6 GB 20K 111 −5.4% ppl Long context
Q2_K 1.6 GB 3.3 GB 22K 130 −15% ppl Long context

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
$ ollama pull qwen3:4b
$ OLLAMA_CONTEXT_LENGTH=8192 \
    ollama run qwen3:4b

The default. One binary, a model registry, an OpenAI-compatible port. More on Ollama.

01Download is 2.3 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.
03There is room to go to 17K context on this card.
See all models for this rig Compare with another model

Related models on a GeForce GTX 1660 Super

ModelTotalTok/s est.
Qwen3 8B
8.19B · same family
Q4_K_M · 6.3 GB ~22 CPU offload
Qwen3 1.7B
1.72B · same family
Q4_K_M · 2.4 GB 210 Runs great
Qwen3 14B
14.8B · same family
Q4_K_M · 10.2 GB ~6.6 CPU offload
Qwen3 0.6B
0.6B · same family
Q4_K_M · 1.8 GB 603 Runs great
Ministral 3 3B
3.85B · similar size
Q4_K_M · 3.6 GB 94 Runs great
Phi-4-mini 3.8B
3.84B · similar size
Q4_K_M · 3.8 GB 94 Runs great
Gemma 3 4B
4.3B · similar size
Q4_K_M · 3.3 GB 84 Runs great

Every model against the GeForce GTX 1660 Super →

Qwen3 4B on other hardware

DeviceTotalTok/s est.
GeForce RTX 3060 Laptop
5.3 GB usable · 336 GB/s
4.0 GB 90 Runs great
GeForce RTX 4060 Laptop
7.0 GB usable · 256 GB/s
4.0 GB 69 Runs great
M3 · 8 GB
5.4 GB usable · 100 GB/s
4.0 GB 25 Runs great
GeForce RTX 3060 12 GB
10.6 GB usable · 360 GB/s
4.0 GB 96 Runs great
M4 · 16 GB
10.7 GB usable · 120 GB/s
4.0 GB 30 Runs great
GeForce RTX 4090
22.4 GB usable · 1008 GB/s
4.0 GB 270 Runs great
Radeon RX 7900 XTX
22.4 GB usable · 960 GB/s
4.0 GB 257 Runs great
M3 Max · 64 GB
48.0 GB usable · 400 GB/s
4.0 GB 99 Runs great

All 114 devices, by tier →