How much VRAM does Gemma 4 31B need?

About 22.8 GB at Q4_K_M, 39.2 GB at Q8_0 and 72.9 GB at FP16, estimated for a ~4k context.

How fast is Gemma 4 31B on an RTX 3090?

Our estimate is 30–42 tokens per second at Q4_K_M. It is not a measured benchmark.

Home Can I run it?Gemma 4 31BGemma 4 31B on an RTX 3090

Yes, but it is tight. Gemma 4 31B at Q4_K_M needs about 22.8 GB, and the GeForce RTX 3090 has 24.0 GB usable at a ~4k context.

Can I run Gemma 4 31B on an RTX 3090?

Gemma 4 31B at Q4_K_M needs about 22.8 GB, and the GeForce RTX 3090 has 24.0 GB usable. That figure is 24 GB of VRAM on the spec sheet, of which 24.0 GB is usable for weights and cache, at 936 GB/s memory bandwidth. Google’s dense Gemma 4 31B. Close in memory to Qwen3-32B, with a different training mix. These are estimates from our formula, not measured benchmarks.

Quant that fits: Q4_K_M Fits, short context only
Memory needed: 22.8 GB at Q4_K_M, ~4k context
Usable VRAM: 24.0 GB of 24 GB VRAM
Est. tokens/sec: 30–42 estimate

Gemma 4 31B · GeForce RTX 3090 · Everything this machine runs · Q4 vs Q8 vs FP16

Memory at each quant

Quant	Need vs 24.0 GB	Verdict
Q4_K_M	22.8 GB	Q4Tight
Q8_0	39.2 GB	Q8Won’t fit
FP16	72.9 GB	FP16Won’t fit

Estimated tokens per second

Estimate: 30–42 tokens/sec at Q4_K_M. This is a bandwidth-derived range, not a bench. Efficiency is a 60–85% GPU band from How we calculate.

Context headroom

We do not have a published KV config for this model, so the memory need is the ~4k-context estimate only. Use the VRAM calculator to explore longer chats.

What to do instead

Run Qwen3-30B-A3B instead (22.3 GB at Q4). It is a different model, not a drop-in, but it does fit this machine.
Q8_0 needs 39.2 GB and does not fit here. The Mac mini M5 Pro 64GB has 48.0 GB usable, enough for Q8.
See the best GPU for this model family and the Gemma 4 31B model page.

Other hardware for Gemma 4 31B

Gemma 4 31B on an RTX 3060 12GBWon’t fit
Gemma 4 31B on an RTX 3070 TiWon’t fit
Gemma 4 31B on an RTX 3080Won’t fit
Gemma 4 31B on an RTX 4060 Ti 16GBWon’t fit
Gemma 4 31B on an RTX 4070 Ti SuperWon’t fit
Gemma 4 31B on an RTX 4090Tight

Other models on an RTX 3090

Quick answers

Can I run Gemma 4 31B on an RTX 3090?: Yes, but it is tight. Gemma 4 31B at Q4_K_M needs about 22.8 GB, and the GeForce RTX 3090 has 24.0 GB usable.
How much VRAM does Gemma 4 31B need?: About 22.8 GB at Q4_K_M, 39.2 GB at Q8_0 and 72.9 GB at FP16, estimated for a ~4k context.
How fast is Gemma 4 31B on an RTX 3090?: Our estimate is 30–42 tokens per second at Q4_K_M. It is not a measured benchmark.