Rent a GPU

Can the NVIDIA RTX 5090 run Gemma 3 12B?

RTX 5090 × Gemma 3 12B8K context
YesRuns fully on the GPU
Memory at FP1628.2 GB of 32 GB
Weights 24.4 GBKV cache 3.2 GBRuntime 0.6 GB
QuantFP16Full quality
Memory28.2 GBof 32 GB usable
Speed~44tokens per second
  1. FP16
  2. Q8_0
  3. Q6_K
  4. Q5_K_M
  5. Q4_K_M
  6. Q3_K_M
  7. Q2_K

Yes. The NVIDIA RTX 5090 can run Gemma 3 12B fully in VRAM at FP16 (full quality), using about 28.2 GB of the 32 GB available, at roughly 44 tokens/second.

Try other context lengths

Run it with Ollama

With Ollama installed, one command downloads and starts Gemma 3 12B:

ollama run gemma3:12b

Ollama's default download is Q4_K_M (about 11.2 GB here). You have room for FP16, which is full quality: pick that GGUF in LM Studio or from Hugging Face.

Run it in the cloud

No NVIDIA RTX 5090? Run Gemma 3 12B in the cloud.

Rent a cloud GPU by the hour. No upfront cost, stop any time.

Rent a GPU on RunPodRent a GPU on Vast.ai

Memory needed by quantization

At 8,192 tokens of context. The black line marks the 32 GB the NVIDIA RTX 5090 can use.

QuantQualityMemoryFits?Speed
FP16Full quality28.2 GBFits~44 tok/s
Q8_0Near-lossless16.8 GBFits~83 tok/s
Q6_KExcellent13.8 GBFits~107 tok/s
Q5_K_MVery good12.5 GBFits~124 tok/s
Q4_K_MGood, most popular11.2 GBFits~145 tok/s
Q3_K_MNoticeably degraded9.8 GBFits~180 tok/s
Q2_KHeavily degraded8.4 GBFits~235 tok/s

GPUs that run Gemma 3 12B comfortably

Models that run well on the NVIDIA RTX 5090

Frequently asked questions

Can NVIDIA RTX 5090 run Gemma 3 12B?

Yes. The NVIDIA RTX 5090 can run Gemma 3 12B fully in VRAM at FP16 (full quality), using about 28.2 GB of the 32 GB available, at roughly 44 tokens/second.

How much VRAM does Gemma 3 12B need?

About 11.2 GB at Q4_K_M with 8,192 tokens of context, and about 28.2 GB at full FP16 precision.

How fast is Gemma 3 12B on the NVIDIA RTX 5090?

Roughly 44 tokens per second at FP16, estimated from the NVIDIA RTX 5090's 1792 GB/s memory bandwidth.