Yes — with 41.0 GB to spare

Phi-4-mini 3.8B at Q4_K_M fits your 2× GeForce RTX 3090 entirely on the GPUs at 8K context, at an estimated 262 tokens per second. There is room for its full 128K window.

Fully on GPU 8K context Q4_K_M · 2.2 GB MIT Released Feb 2025

MIT-licensed, dense, and unusually strong on instruction following for its size.

Hardware requirements What hardware do I need to run Phi-4-mini 3.8B? →

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 2.2 GB
Weights 2.2 GB KV cache @ 8K 1.00 GB Runtime overhead 0.6 GB Free 41.0 GB of 44.8 GB

Quantisation ladder

QuantWeightsTotal @ 8KMax contextTok/sQualityFit
F16 7.2 GB 8.8 GB 128K 79 Reference Long context
Q8_0 3.8 GB 5.4 GB 128K 149 −0.1% ppl Long context
Q6_K 2.9 GB 4.5 GB 128K 193 −0.4% ppl Long context
Q5_K_M 2.5 GB 4.1 GB 128K 224 −0.8% ppl Long context
Q4_K_M 2.2 GB 3.8 GB 128K 262 −1.9% ppl Recommended
Q3_K_M 1.7 GB 3.3 GB 128K 324 −5.4% ppl Long context
Q2_K 1.5 GB 3.1 GB 128K 378 −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 phi4-mini:3.8b
$ OLLAMA_CONTEXT_LENGTH=8192 \
    ollama run phi4-mini:3.8b

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

01Download is 2.2 GB. Keep it on an SSD — a first load off a spinning disk takes minutes.
02llama.cpp and Ollama spread the layers over both cards on their own. The cards take turns on every token, so expect about one card's speed; vLLM's --tensor-parallel-size 2 runs them in parallel instead.
03There is room to go to the model's full 128K context on this card.
See all models for this rig Compare with another model

Related models on a 2× GeForce RTX 3090

ModelTotalTok/s est.
Phi-4 14B
14.7B · same family
Q4_K_M · 10.4 GB 69 Runs great
Ministral 3 3B
3.85B · similar size
Q4_K_M · 3.6 GB 262 Runs great
Qwen3 4B
4.02B · similar size
Q4_K_M · 4.0 GB 251 Runs great
Gemma 3 4B
4.3B · similar size
Q4_K_M · 3.3 GB 234 Runs great
Granite 4.1 3B
3.4B · similar size
Q4_K_M · 3.1 GB 296 Runs great
Llama 3.2 3B Instruct
3.21B · similar size
Q4_K_M · 3.3 GB 314 Runs great
Qwen3.5 4B
4.66B · similar size
Q4_K_M · 3.5 GB 216 Runs great

Every model against the 2× GeForce RTX 3090 →

Phi-4-mini 3.8B on other hardware

DeviceTotalTok/s est.
GeForce RTX 3060 Laptop
5.3 GB usable · 336 GB/s
3.8 GB 94 Runs great
GeForce RTX 4060 Laptop
7.0 GB usable · 256 GB/s
3.8 GB 72 Runs great
M3 · 8 GB
5.4 GB usable · 100 GB/s
3.8 GB 26 Runs great
GeForce RTX 3060 12 GB
10.6 GB usable · 360 GB/s
3.8 GB 101 Runs great
M4 · 16 GB
10.7 GB usable · 120 GB/s
3.8 GB 31 Runs great
GeForce RTX 4090
22.4 GB usable · 1008 GB/s
3.8 GB 283 Runs great
Radeon RX 7900 XTX
22.4 GB usable · 960 GB/s
3.8 GB 269 Runs great
M3 Max · 64 GB
48.0 GB usable · 400 GB/s
3.8 GB 104 Runs great

All 114 devices, by tier →