CalcuGrid

What LLMs Can I Run on a GeForce RTX 3090?

24 GB of VRAM, of which about 22.1 GB is usable for a model. That runs 16 of the 20 models here , up to Qwen3 32B.

Largest model that fits: Qwen3 32B at Q4_K_M, needing 20.0 GB and leaving 2.1 GB spare. Its context can run to 17K tokens before memory runs out. The value choice for local inference: 24 GB at a fraction of the price of a current card, which is why paired 3090s are the commonest home build for large models.

Every model, on this hardware

Which models run on a GeForce RTX 3090
Model Size Needs Runs Max context
Llama 3.2 1B 1.2B 1.7 GB Yes 128K
Llama 3.2 3B 3.2B 3.4 GB Yes 128K
Llama 3.1 8B 8.0B 6.0 GB Yes 128K
Llama 3.3 70B 70.5B 40.3 GB No —
Qwen2.5 3B 3.1B 2.7 GB Yes 32K
Qwen2.5 7B 7.6B 5.2 GB Yes 32K
Qwen2.5 14B 14.8B 10.0 GB Yes 32K
Qwen2.5 32B 32.8B 20.0 GB Yes 17K
Qwen2.5 72B 72.7B 41.4 GB No —
Qwen2.5 Coder 7B 7.6B 5.2 GB Yes 32K
Qwen2.5 Coder 32B 32.8B 20.0 GB Yes 17K
Qwen3 32B 32.8B 20.0 GB Yes 17K
Mistral 7B v0.3 7.3B 5.6 GB Yes 32K
Gemma 2 9B 9.2B 8.3 GB Yes 8K
Gemma 2 27B 27.2B 17.9 GB Yes 8K
Phi-4 14.7B 10.0 GB Yes 16K
Qwen3 30B-A3B 30.5B 17.5 GB Yes 40K
Qwen3 235B-A22B 235.1B 125.4 GB No —
Mixtral 8x7B 46.7B 26.3 GB No —
DeepSeek-R1 Distill Qwen 14B 14.8B 10.0 GB Yes 72K

Q4_K_M at 8K context. "Max context" is the longest context that still fits once the weights are loaded.

Memory decides this, not speed

For inference, memory capacity decides what runs and bandwidth decides how fast. A slower card with more memory beats a faster card with less, every time, because a model that does not fit does not run at any speed. That is why a 16 GB RTX 4060 Ti is a more useful local-LLM card than a faster 12 GB one.

The other half is context. Weights are a fixed cost, but the key-value cache grows with every token, and on a long conversation it can outgrow the weights. That is why the table above gives a maximum context per model rather than a yes or no — plenty of models load fine on this hardware and then run out of room mid-conversation.

Frequently asked questions

What LLMs can a GeForce RTX 3090 run?
16 of the 20 models on this site, at Q4_K_M with a 8K context. The largest is Qwen3 32B at 32.8B parameters, needing 20.0 GB of the roughly 22.1 GB usable.
Why is usable memory less than the 24 GB on the box?
The framework, the driver and, on a card driving a display, the screen itself all take a share before the weights get any. This page assumes about 22.1 GB of the 24 GB is available, which is why a model that looks like it fits on paper sometimes does not.
Does a bigger model at a rougher quantisation beat a smaller one?
Usually, yes. A 32B model at Q3_K_M generally outperforms a 7B at Q8_0, even though the smaller model is barely quantised at all. The practical rule is to fit the largest model you can and then step the quantisation down only as far as you must — which is what the "best format" column above is showing.
How accurate are these figures?
They count weights, the key-value cache at the chosen context, and a fixed runtime allowance. Real usage moves with the runtime, the batch size and whether flash attention is enabled. Treat anything within a gigabyte of capacity as "probably not" rather than "just fits" — which is why the table says "only just" rather than "yes" past 95% of usable memory.

Manufacturer specifications for VRAM and unified memory. Model config.json files published on Hugging Face.

An estimate of memory, not a benchmark. It counts model weights, the key-value cache at the context you choose, and a fixed runtime allowance. Actual usage moves with the runtime, the batch size, whether flash attention is on, and how much the operating system has already taken from a shared memory pool. Treat a result within a gigabyte of your card's capacity as 'probably not' rather than 'just fits'.

Data on this page last verified .

    ↑↓ to move ↵ to open