Gemma31B params128k contextpopular
Gemma 4 31B locally
Gemma 4 31B is an open-weight LLM from the Gemma family with 31B parameters. Main use: chat, RAG and general assistance. Detected minimum hardware: RTX A4500 (20 GB).
Technical facts
Parameters31B
Max context128k
Q4_K_M19.5 GB
Q5_K_M23.8 GB
Q834.6 GB
FP1669.3 GB
FamilyGemma
Last sync2026-05-12
Available quantizations
GGUF weightsQ4_K_M
19.5GB
Acceptable. Good compromise when VRAM is limited.
Q5_K_M
23.8GB
Good quality. Sweet spot for size and precision.
Q8
34.6GB
Near-FP16 quality. Comfortable for production.
FP16
69.3GB
Reference precision. Maximum quality, doubled VRAM.
Compatible GPUs
12 single-GPUGPUs that can run Gemma 4 31B on a single card, ranked by VRAM headroom.
RTX A4500
NVIDIA20 GB · RTX A (Ampere)
19.5 / 20 GBtight · Q4
RTX 4000 Ada
NVIDIA20 GB · RTX Ada
19.5 / 20 GBtight · Q4
Radeon RX 7900 XT
AMD20 GB · RDNA 3
19.5 / 20 GBtight · Q4
TITAN RTX
NVIDIA24 GB · RTX 20
19.5 / 24 GBcomfortable · Q4
RTX 3090
NVIDIA24 GB · RTX 30
19.5 / 24 GBcomfortable · Q4
RTX 3090 Ti
NVIDIA24 GB · RTX 30
19.5 / 24 GBcomfortable · Q4
RTX 4090
NVIDIA24 GB · RTX 40
19.5 / 24 GBcomfortable · Q4
Quadro RTX 6000
NVIDIA24 GB · Quadro RTX
19.5 / 24 GBcomfortable · Q4
RTX A5000
NVIDIA24 GB · RTX A (Ampere)
19.5 / 24 GBcomfortable · Q4
RTX A5500
NVIDIA24 GB · RTX A (Ampere)
19.5 / 24 GBcomfortable · Q4
RTX 4500 Ada
NVIDIA24 GB · RTX Ada
19.5 / 24 GBcomfortable · Q4
NVIDIA A10
NVIDIA24 GB · Ampere DC
19.5 / 24 GBcomfortable · Q4
Recommended multi-GPU rigs
2x / 4x consumer GPUsFor Gemma 4 31B at higher quantization or with more context, a multi-GPU rig gives more headroom.
2× RTX 3080 10GB
NVIDIA20 GB · RTX 30
19.5 / 20 GBtight · Q4
2× Radeon RX 6700
AMD20 GB · RDNA 2
19.5 / 20 GBtight · Q4
2× Arc B570
Intel20 GB · Arc Battlemage
19.5 / 20 GBtight · Q4
2× GTX 1080 Ti
NVIDIA22 GB · GTX 10
19.5 / 22 GBtight · Q4
2× RTX 2080 Ti
NVIDIA22 GB · RTX 20
19.5 / 22 GBtight · Q4
4× GTX 1060 6GB
NVIDIA24 GB · GTX 10
19.5 / 24 GBcomfortable · Q4
4× GTX 1660
NVIDIA24 GB · GTX 16
19.5 / 24 GBcomfortable · Q4
4× GTX 1660 Super
NVIDIA24 GB · GTX 16
19.5 / 24 GBcomfortable · Q4
Recommended rig
4× GTX 1060 6GB
Gemma 4 31B with Ubuntu, vLLM, Open WebUI and the model already downloaded.
Similar models
VRAM estimates: parameters x bits/8 plus margin. Real performance varies by engine, context length and batch size.
sync: 2026-05-12