Can you run Gemma 3 27B locally?

Gemma 3 family · March 2025 · 27.40B parameters · Hugging Face model card

How much VRAM does Gemma 3 27B need at each quantization?

Gemma 3 27B needs 20.2 GB of VRAM at Q4_K_M. The table below lists weights-only size and total VRAM including overhead for each common quantization level with 4k and 32k token contexts.

Gemma 3 uses a 1024-token sliding attention window on 5 of every 6 layers; the table shows the conservative full-attention upper bound, so real long-context use is slightly lower.

Method: weights = parameters × bits-per-weight (Q4_K_M ≈ 4.8, Q5_K_M ≈ 5.7, Q6_K ≈ 6.6, Q8_0 ≈ 8.5, FP16 = 16), plus 10% loading overhead, plus KV cache from the verified architecture config (62 layers, 16 KV heads, 128 head dim). Source: Params 27.4B, 62 layers, 16 KV heads, head dim 128, 131072 context, sliding window 1024 — HF config.json (unsloth/gemma-3-27b-it mirror of google/gemma-3-27b-it).

Which GPUs can run Gemma 3 27B locally?

At Q4_K_M with a 4k context, Gemma 3 27B needs 20.2 GB of VRAM. The lists below are computed live from our GPU database and grouped by how much headroom the card has. Disclosure: CompareAIHardware.com participates in the Amazon Associates program and earns from qualifying purchases through links on this page. Affiliate relationships do not influence our recommendations.

Runs Gemma 3 27B comfortably (25%+ VRAM headroom)

Minimum GPUs that fit Gemma 3 27B at Q4_K_M

These GPUs hold the model but leave little headroom — keep contexts short.

Needs 2+ GPUs to run Gemma 3 27B

One of these cards is too small on its own, but a pair (tensor or pipeline parallel, ~90% efficiency) covers the 20.2 GB requirement. See our multi-GPU guide for setup.

Can a Mac run Gemma 3 27B?

Yes — these Apple Silicon machines fit Gemma 3 27B at Q4_K_M, since macOS lets the GPU use about 75% of unified memory. Generation speed is bound by memory bandwidth, so the GB/s column matters as much as capacity.

Can Gemma 3 27B run on mini PCs, Jetson, or NPU devices?

These edge and NPU devices from our database have enough memory for Gemma 3 27B at Q4_K_M. Their memory bandwidth is far below discrete GPUs, so expect a fraction of desktop generation speed.

Which pre-built systems can run Gemma 3 27B?

These mini PCs and workstations from our database fit Gemma 3 27B at Q4_K_M. Usable-memory figures are conservative: Windows shares about half of system RAM with the GPU by default (Linux can expose more), macOS lets Apple Silicon GPUs use about 75% of unified memory, and Linux unified-memory systems such as GB10 expose roughly 90%. Generation speed is bound by memory bandwidth, so compare the GB/s column before buying.

Frequently asked questions

Related guides

• Best GPUs for LLMs
• VRAM Calculator
• Best GPU for Stable Diffusion