Skip to content
Run LLM Locally

Gemma 2 9B IT on the RTX 5090

With 32 GB of VRAM, the RTX 5090 runs Gemma 2 9B IT at Q4_K_M entirely on the GPU with an 8K context (about 8.7 GB needed). At Q8_0 it fits entirely.

ShareFacebookXWhatsAppLinkedInEmail

Gemma 2 9B IT

Q4_K_M · 8K context

Excellent

The model needs about 8.7 GB and fits in the 31 GB available, with 22.3 GB to spare.

Your hardware
RTX 5090 32 GB · 64 GB RAM
Estimated memory
~8.7 GB estimate
VRAM (usable memory)
31 GB
GPU only
Yes
CPU/GPU offload
Not required
Generation speed
134–213 tok/s estimate
Largest context in fast memory
8K tokens
Confidence: low

Based on: memory bandwidth (1,792 GB/s) and model size

Memory breakdown

~8.7 GB (8.2 GB–9.3 GB) needed · estimate

Fits entirely in VRAM · 22.3 GB left

  • Model weights5.4 GB
  • KV cache2 GB
  • Runtime overhead1 GB
  • Safety margin0.4 GB
Assumptions: context 8K tokens · KV cache f16 · llama.cpp (CUDA) · flash attention on · weights = real file size · sliding-window layers keep 4,096 tokens

EstimateData updated 24 days agoSource: Run LLM Locally engine, Hugging Face config.json + model card

What makes this estimate less reliable

  • no benchmark for this hardware and model yet

Other quantizations on the RTX 5090

Gemma 2 9B IT on RTX 5090 + 64 GB RAM · 8K context
Gemma 2 9B IT Q8_08K~12.7 GB · 96–143 tok/s estimateExcellent
Gemma 2 9B IT Q6_K8K~10.5 GB · 114–175 tok/s estimateExcellent
Gemma 2 9B IT Q5_K_M8K~9.6 GB · 124–193 tok/s estimateExcellent
Gemma 2 9B IT Q4_K_M8K~8.7 GB · 134–213 tok/s estimateExcellent
Gemma 2 9B IT IQ4_XS8K~8.2 GB · 142–229 tok/s estimateExcellent
Gemma 2 9B IT Q3_K_M8K~7.8 GB · 149–242 tok/s estimateExcellent

Alternatives

GPUs that run Gemma 2 9B IT Q4_K_M entirely in VRAM

Other models that fit the RTX 5090

Where to buy a RTX 5090

Prices and links follow the market you choose.

GeForce RTX 5090

32 GB GDDR7 · 1,792 GB/s · 575 W

ExcellentRuns Gemma 2 9B IT Q4_K_M entirely in memory

Price not available

Availability unknown

View on amazon.com

Affiliate link

No current offer imported for this market: the link opens a marketplace search.

Frequently asked questions

Can the RTX 5090 run Gemma 2 9B IT?
The model needs about 8.7 GB and fits in the 31 GB available, with 22.3 GB to spare.
Which quantization should I use on the RTX 5090?
Q8_0 is the highest-precision quantization that fits entirely in VRAM at 8K context. ~12.7 GB / 31 GB.
What is the largest context on the RTX 5090?
At Q4_K_M, about 8K tokens fit entirely in VRAM with an f16 KV cache. A q8_0 KV cache roughly doubles that.
How fast is Gemma 2 9B IT (RTX 5090)?
We estimate 134–213 tok/s (estimate, confidence low). Based on: memory bandwidth (1,792 GB/s) and model size.