Can the NVIDIA RTX 4080 run Gemma 2 27B?

Yes, with heavy quantization
QuantQ2_KHeavily degraded
Memory13.9 GBof 16 GB
Speed~42tokens/s

Only just. Gemma 2 27B fits on the NVIDIA RTX 4080 at Q2_K, which costs noticeable quality. Expect about 42 tokens/second. A card with more memory is a better experience.

Try other context lengths in the calculator →
Run it in the cloud

Want full quality? Run Gemma 2 27B on a bigger rented GPU.

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

RunPodOn-demand GPUs with ready-made templates for Ollama and vLLM.Rent a GPU on RunPod → Vast.aiA marketplace of rentable GPUs, from RTX cards to H100s.Rent a GPU on Vast.ai →

Memory needed by quantization

At 8,192 tokens of context.

QuantQualityMemoryFits?Speed
FP16Full quality58.1 GBToo big—
Q8_0Near-lossless32.6 GBToo big—
Q6_KExcellent26 GBToo big—
Q5_K_MVery good23 GBToo big—
Q4_K_MGood, most popular20.2 GBToo big—
Q3_K_MNoticeably degraded17 GBToo big—
Q2_KHeavily degraded13.9 GBFits~42 tok/s

How to run Gemma 2 27B on the NVIDIA RTX 4080

Ollama's default gemma2:27b download is Q4_K_M, which needs about 20.2 GB, more than the 16 GB available. Download the Q2_K GGUF instead (LM Studio lists quants per model), or try a smaller context length.

GPUs that run Gemma 2 27B comfortably

Check NVIDIA RTX 3090 prices →

Models that run well on the NVIDIA RTX 4080