Skip to content
Llama Builds

Google · Released March 2025

Gemma 3 27B

What it takes to run Gemma 3 27B locally: memory by quantization, the GPUs that handle it, and how fast it runs.

Updated

Parameters
27.4B
Architecture
Dense
Context
128K tokens
License
Gemma Terms of Use
Memory at Q4_K_M
~19GB

What do you need to run Gemma 3 27B?

Gemma 3 27B needs about 19GB of VRAM at Q4_K_M with 8K context (17GB for the weights plus KV cache and overhead). The smallest GPU that can hold it is a 16GB card like the RTX 4060 Ti 16GB at Q3_K_M, with little memory to spare. For a good balance of price and speed we recommend the RTX 3090 (~39 tokens/sec).

See the full hardware guide for Gemma 3 27B

What to buy

GPUs for Gemma 3 27B

Cheapest that runs it

RTX 4060 Ti 16GB

16GB · ~$380 used

Quant
Q3_K_M
Speed (est.)
~17 tok/s
Fits, tight
Check price at Amazon

Best value

RTX 3090

24GB · ~$1,050 used

Quant
Q4_K_M
Speed (est.)
~39 tok/s
Runs well
Check price at Amazon

Fastest

RTX 5090

32GB · ~$4,300

Quant
Q6_K
Speed (est.)
~60 tok/s
Runs well
Check price at Amazon

Speeds are estimates at the quantization shown with 8K context. We may earn a commission when you buy through these links. Learn more.

Memory

Gemma 3 27B VRAM requirements by quantization

Lower-bit quantizations shrink the model at some cost to quality. Q4_K_M is the usual sweet spot. Totals include the KV cache for the context length shown plus runtime overhead.

Memory needed for Gemma 3 27B at each quantization level
QuantizationQualityWeightsTotal @ 8K ctxTotal @ 32K ctx
Q2_KRough10 GB12 GB14 GB
Q3_K_MUsable13 GB15 GB17 GB
Q4_K_MRecommendedDefault17 GB19 GB21 GB
Q5_K_MVery good19 GB22 GB24 GB
Q6_KNear-lossless23 GB25 GB27 GB
Q8_0Practically lossless29 GB31 GB33 GB
BF16Full precision54 GB57 GB59 GB

KV cache note: 62 layers in a 5:1 pattern of 1,024-token sliding-window and global layers. Only the 10 global layers grow with context: 2 x 10 x 16 KV heads x 128 dims x 2 bytes = 81,920 bytes per token, about 10.7GB at the full 128K. The 52 sliding layers add a fixed buffer of about 0.4GB.

Performance

How fast Gemma 3 27B runs

Single-user generation speed on popular GPUs that fit the model entirely in memory, at the best quantization each can hold.

  1. RTX 5090

    Q6_K · 32GB

    ~60tok/s

  2. RTX 5080

    Q3_K_M · 16GB

    ~55tok/s

  3. RTX 5070 Ti

    Q3_K_M · 16GB

    ~51tok/s

  4. RTX PRO 6000

    Q8_0 · 96GB

    ~48tok/s

  5. RTX 4090

    Q4_K_M · 24GB

    ~46tok/s

  6. RTX 4080 Super

    Q3_K_M · 16GB

    ~42tok/s

  7. RTX 3090

    Q4_K_M · 24GB

    ~39tok/s

  8. RTX 4070 Ti Super

    Q3_K_M · 16GB

    ~39tok/s

  9. RTX PRO 5000 72GB

    Q8_0 · 72GB

    ~36tok/s

  10. RX 7900 XTX

    Q4_K_M · 24GB

    ~36tok/s

Estimated (method)

Overview

About Gemma 3 27B

Gemma 3 27B was one of the most popular 24GB-GPU models of 2025, praised for natural writing, image understanding, and broad multilingual support. Gemma 4 31B replaces it with a reasoning mode, a 256K context, much stronger coding and math, and an Apache 2.0 license, so it is now superseded. At Q4_K_M it needs about 18GB with 8K context, leaving room for long context on a 24GB card, and Google's quantization-aware Q4_0 release keeps quality close to the original at a similar size. Its sliding-window design keeps the KV cache small, about 0.7GB per 8K tokens.

Strengths

  • +Natural, fluent writing and strong multilingual chat
  • +Image input included
  • +Small long-context memory cost thanks to sliding-window attention
  • +Official quantization-aware (QAT) 4-bit weights from Google

Weaknesses

  • −No reasoning mode; weak at math and complex coding
  • −Superseded by Gemma 4 31B and Gemma 4 26B A4B
  • −Gemma license is more restrictive than Apache 2.0
ChatVisionMultilingualWritingImage input

Setup

How to run Gemma 3 27B

Ollama
$ ollama run gemma3:27b
llama.cpp
$ llama-server -hf bartowski/google_gemma-3-27b-it-GGUF:Q4_K_M -c 8192 -ngl 99
MLX (Mac)
$ mlx_lm.chat --model mlx-community/gemma-3-27b-it-4bit

Weights: google/gemma-3-27b-it · GGUF: bartowski/google_gemma-3-27b-it-GGUF

Alternatives

Models like Gemma 3 27B

Google

Gemma 4 31B

31B · 256K ctx

~22GBat Q4_K_M
ChatVisionReasoning

Google

Gemma 4 26B A4B

26B (3.8B active) · 256K ctx

~17GBat Q4_K_M
ChatVisionReasoning

Mistral AI

Mistral Small 3.2 24B Instruct

24B · 128K ctx

~17GBat Q4_K_M
ChatVisionAgents & tools

Google

Gemma 3 12B

12.2B · 128K ctx

~9GBat Q4_K_M
ChatVisionMultilingual

Frequently asked questions

How much VRAM do you need to run Gemma 3 27B?

About 19GB at Q4_K_M with 8K tokens of context: 16.6GB of weights, 1.1GB of KV cache, and 1.1GB of runtime overhead. Higher-precision quants and longer contexts need more; see the table above.

What is the cheapest GPU that can run Gemma 3 27B?

The smallest card in our database that can hold it is the RTX 4060 Ti 16GB (16GB), at Q3_K_M with little memory to spare, generating ~17 tokens/sec.

Is Gemma 3 27B free for commercial use?

Gemma 3 27B is released under the Gemma Terms of Use. Commercial use is allowed with some conditions; read the license before shipping a product.

How long a context does Gemma 3 27B support?

Up to 131,072 tokens (128K). Long contexts need a lot of extra memory for the KV cache, so most local setups run 8K to 32K.

Sources: Gemma 3 27B IT model card (Hugging Face), Gemma 3 technical report. Capability score basis: Placed between the Llama 3.1 8B and Qwen3 32B anchors from its Artificial Analysis Intelligence Index v3 standing (October 2025), adjusted up slightly for chat and writing quality.