NV NVIDIA Consumer Discrete GPU

NVIDIA GeForce RTX 3090

Runs downloadable LLMs up to ~32.5B at Q4, ~152.5-181.8 tok/s on an 8B model.

NVIDIA GeForce RTX 3090
24GB
VRAM
936
GB/s bandwidth
32.5B
Max model (Q4)
$1,499
MSRP

What you can run

At 8k context, FP16 KV. Full-speed tok/s holds up to "max ctx"; beyond it the context spills to system RAM (CPU/PCIe offload) and drops to the degraded figure.

Model sizeFP16Q8Q4
3B 113.9-135.7 t/s
≤222k ctx · then ~4.9-5.8 (slow)
227.8-271.4 t/s
≤268k ctx · then ~9.7-11.6 (slow)
406.7-484.7 t/s
≤288k ctx · then ~17.4-20.7 (slow)
7-8B 42.7-50.9 t/s
≤35k ctx · then ~1.8-2.2 (slow)
85.4-101.8 t/s
≤96k ctx · then ~3.7-4.4 (slow)
152.5-181.8 t/s
≤123k ctx · then ~6.5-7.8 (slow)
13B won't run 52.6-62.6 t/s
≤46k ctx · then ~2.2-2.7 (slow)
93.9-111.9 t/s
≤81k ctx · then ~4-4.8 (slow)
32B won't run won't run 38.1-45.4 t/s
≤10k ctx · then ~1.6-1.9 (slow)
70B won't run won't run won't run
120B won't run won't run won't run
400B won't run won't run won't run

Estimate: tok/s ≈ bandwidth ÷ (params × bytes/param) × efficiency (NVIDIA). Cyan = fits ~22.1 GB usable. Beyond "ctx", KV spills to system RAM (CPU/PCIe offload) (rough degraded speed shown).

What else you need for a full build

Minimum supporting components for a desktop built around the NVIDIA GeForce RTX 3090. Guidance — exact needs depend on your case, other parts and how much you offload to system RAM.

Power supply
≥ 750W · 80+ Gold
Headroom over the 350W card plus the rest of the build.
System RAM
≥ 32 GB · 64 GB recommended
At least the 24 GB of VRAM; 2× lets you CPU-offload models larger than VRAM.
Motherboard
1× free PCIe x16 slot (4.0/5.0)
A full-length x16 electrical slot; PCIe 4.0 is ample, 5.0 future-proofs.
CPU
Any modern multi-core
The GPU runs inference; more cores and memory channels only matter when offloading to RAM.
Storage
≥ 1 TB NVMe SSD
Weights are large — a 70B model at Q4 is ~40 GB; budget for a few.
PSU connector
12V-2×6 / 12VHPWR (or 3× 8-pin)
High-power NVIDIA cards use the 16-pin 12V-2×6 power connector.

Run popular models on the NVIDIA GeForce RTX 3090

Videos about the NVIDIA GeForce RTX 3090

RTX 3090: Still the Best Budget AI GPU?

Local Ai Server Benchmark 3090 vs Dual 3060s Performance is INSANE!

More hardware like the NVIDIA GeForce RTX 3090

About the NVIDIA GeForce RTX 3090

NVIDIA GeForce RTX 3090 is a discrete GPU from NVIDIA on Ampere, 2020. For local genAI the numbers that matter are its 24 GB (how big a model + context fits) and 936 GB/s bandwidth (generation speed). Plan for a ~750W power supply.

Frequently asked questions

What size LLM can the NVIDIA GeForce RTX 3090 run?

With ~22.1 GB usable it runs up to ~32.5B at Q4 (~9.5B at FP16) at 8k context. Bigger spills into system RAM (CPU/PCIe offload) and slows sharply.

How many tokens/sec does the NVIDIA GeForce RTX 3090 generate?

Generation is bandwidth-bound. At 936 GB/s an 8B model at Q4 runs ~152.5-181.8 tok/s (estimate).

Why does long context slow it down?

The KV cache for the context must also fit fast memory; once weights+KV exceed 22.1 GB it spills to system RAM (CPU/PCIe offload) and throughput collapses. Each row shows the max usable context.

What else do I need to run the NVIDIA GeForce RTX 3090?

Besides the card: a ≥750W 80+ Gold power supply, at least 24GB of system RAM (2× the 24GB VRAM is ideal for CPU offload), a free PCIe x16 slot, any modern multi-core CPU, and a ≥1TB NVMe SSD for model weights.