Models / Gemma
Can an RTX 4090 run Gemma 4 31B?
Dense Gemma 4, 30.7B parameters. Q4 is a 24 GB card; Q8 is not.
- Stored parameters
- 30.7B
- Active parameters
- Dense
- Q4 total (estimate)
- 19.4 GB
- License
- See model card
Model card. Google's Gemma 4 card lists the 31B dense model at 30.7B parameters.
Compatibility
| Hardware | Q4 | Q8 | FP16 |
|---|---|---|---|
| Ryzen AI Max+ 395 128GB AMD · 128 GB unified
| 62.1 GB Fits estimate No measured run | ||
| Ryzen AI Max+ 395 64GB AMD · 64 GB unified | 33.4 GB Fits estimate No measured run | 62.1 GB Tight estimate No measured run | |
| RTX 5090 NVIDIA · 32 GB VRAM | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run | |
| RTX 3090 NVIDIA · 24 GB VRAM | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run | |
| M3 Ultra 512GB Apple · 512 GB unified | 19.4 GB Fits estimate No measured run | 33.4 GB Fits estimate No measured run | 62.1 GB Fits estimate No measured run |
| 2× NVIDIA GB10 NVIDIA · 256 GB unified | 19.4 GB Fits estimate No measured run | 33.4 GB Fits estimate No measured run | 62.1 GB Fits estimate No measured run |
| M3 Ultra 256GB Apple · 256 GB unified | 19.4 GB Fits estimate No measured run | 33.4 GB Fits estimate No measured run | 62.1 GB Fits estimate No measured run |
| M5 Ultra 256GB Apple · 256 GB unified | 19.4 GB Fits estimate No measured run | 33.4 GB Fits estimate No measured run | 62.1 GB Fits estimate No measured run |
| M2 Ultra 192GB Apple · 192 GB unified | 19.4 GB Fits estimate No measured run | 33.4 GB Fits estimate No measured run | 62.1 GB Fits estimate No measured run |
| NVIDIA GB10 NVIDIA · 128 GB unified | 19.4 GB Fits estimate No measured run | 33.4 GB Fits estimate No measured run | 62.1 GB Fits estimate No measured run |
| M1 Ultra 128GB Apple · 128 GB unified | 19.4 GB Fits estimate No measured run | 33.4 GB Fits estimate No measured run | 62.1 GB Fits estimate No measured run |
| M5 Max 128GB Apple · 128 GB unified | 19.4 GB Fits estimate No measured run | 33.4 GB Fits estimate No measured run | 62.1 GB Fits estimate No measured run |
| 4× RTX 3090 NVIDIA · 96 GB VRAM | 19.4 GB Fits estimate No measured run | 33.4 GB Fits estimate No measured run | 62.1 GB Fits estimate No measured run |
| RTX Pro 6000 Blackwell NVIDIA · 96 GB VRAM | 19.4 GB Fits estimate No measured run | 33.4 GB Fits estimate No measured run | 62.1 GB Fits estimate No measured run |
| M5 Ultra 96GB Apple · 96 GB unified | 19.4 GB Fits estimate No measured run | 33.4 GB Fits estimate No measured run | 62.1 GB Fits estimate No measured run |
| 2× RTX 3090 NVIDIA · 48 GB VRAM | 19.4 GB Fits estimate No measured run | 33.4 GB Fits estimate No measured run | 62.1 GB Over estimate No measured run |
| M4 Pro 48GB Apple · 48 GB unified | 19.4 GB Fits estimate No measured run | 33.4 GB Fits estimate No measured run | 62.1 GB Over estimate No measured run |
| M4 Max 36GB Apple · 36 GB unified | 19.4 GB Fits estimate No measured run | 33.4 GB Tight estimate No measured run | 62.1 GB Over estimate No measured run |
| M5 Max 36GB Apple · 36 GB unified | 19.4 GB Fits estimate No measured run | 33.4 GB Tight estimate No measured run | 62.1 GB Over estimate No measured run |
| Radeon AI Pro R9700 AMD · 32 GB VRAM | 19.4 GB Fits estimate No measured run | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run |
| Intel Arc Pro B70 Intel · 32 GB VRAM | 19.4 GB Fits estimate No measured run | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run |
| M4 32GB Apple · 32 GB unified | 19.4 GB Fits estimate No measured run | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run |
| M6 32GB Apple · 32 GB unified | 19.4 GB Fits estimate No measured run | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run |
| RTX 4090 NVIDIA · 24 GB VRAM | 19.4 GB Fits estimate No measured run | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run |
| TITAN RTX NVIDIA · 24 GB VRAM | 19.4 GB Fits estimate No measured run | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run |
| Radeon RX 7900 XTX AMD · 24 GB VRAM | 19.4 GB Fits estimate No measured run | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run |
| M4 Pro 24GB Apple · 24 GB unified | 19.4 GB Tight estimate No measured run | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run |
| M3 Pro 18GB Apple · 18 GB unified | 19.4 GB Over estimate No measured run | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run |
| RTX 4060 Ti 16GB NVIDIA · 16 GB VRAM | 19.4 GB Over estimate No measured run | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run |
| RTX 5060 Ti 16GB NVIDIA · 16 GB VRAM | 19.4 GB Over estimate No measured run | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run |
| RTX 5070 Ti NVIDIA · 16 GB VRAM | 19.4 GB Over estimate No measured run | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run |
| M4 16GB Apple · 16 GB unified | 19.4 GB Over estimate No measured run | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run |
| RTX 3060 12GB NVIDIA · 12 GB VRAM | 19.4 GB Over estimate No measured run | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run |
| RTX 3080 Ti NVIDIA · 12 GB VRAM | 19.4 GB Over estimate No measured run | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run |
| RTX 4070 Ti NVIDIA · 12 GB VRAM | 19.4 GB Over estimate No measured run | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run |
| RTX 2060 Super NVIDIA · 8 GB VRAM | 19.4 GB Over estimate No measured run | 33.4 GB Over estimate No measured run | 62.1 GB Over estimate No measured run |
Measured runs
| Hardware | Quant | tok/s | Runtime | Source |
|---|---|---|---|---|
| RTX 5090 | Q4_K_M | 71.3 | llama.cpp b11658 (llama-bench) | Source · checked · prefill 3841 |
| RTX 5090 | llama-bench tg128 (quant per result) | 70.0 | llama.cpp | Source · checked |
| RTX 4090 | llama-bench tg128 (quant per result) | 43.8 | llama.cpp | Source · checked |
| RTX 3090 | Q4_K_M | 38.8 | llama.cpp b11658 (llama-bench) | Source · checked · prefill 1292 |
| Ryzen AI Max+ 395 64GB | q4_0 | 11.28 | llama.cpp (ROCm) | Source · checked · prefill 69.5 |
| Ryzen AI Max+ 395 128GB | Q4_K_M (17.07 GiB) | 11.27 | llama.cpp (Vulkan, commit f5222f25c) | Source · checked · prefill 360 |
| Ryzen AI Max+ 395 128GB | 4-bit | 8.8 | Ollama 0.34.0 | Source · checked |
| Ryzen AI Max+ 395 128GB | Q6_K (23.47 GiB) | 8.1 | llama.cpp (Vulkan, commit f5222f25c) | Source · checked · prefill 348 |
| M6 32GB | QAT (google/gemma-4-31b-qat) | 7.4 | LM Studio | Source · checked |
| Ryzen AI Max+ 395 128GB | Q8_0 (30.39 GiB) | 6.46 | llama.cpp (Vulkan, commit f5222f25c) | Source · checked · prefill 412 |