google · vision language

gemma-3-27b-it-qat-q4_0-unquantized

google/gemma-3-27b-it-qat-q4_0-unquantized

gemma-3-27b-it-qat-q4_0-unquantized at Q4_K_M is exactly 16,546,688,736 bytes (15.41 GiB / 16.55 GB) — an effective 4.825 bits per weight, not the nominal 4.

From the file· summed from 1 file(s)
Parameters
27.4B
Architecture
gemma3
Context
—
native (config.json)
License
gemma

Shipped quantizations

● exact bytes, summed from published files
QuantSize●Exact bytes●Effective bpw●Tensors●Publisher
UD-IQ1_S6.06 GiB6,506,339,4241.897—unsloth
UD-IQ1_M6.51 GiB6,986,738,7842.038—unsloth
IQ2_XXS7.16 GiB7,685,576,0642.241—bartowski
UD-IQ2_XXS7.31 GiB7,850,253,4082.289—unsloth
IQ2_XS7.86 GiB8,438,904,1922.461—bartowski
IQ2_S8.18 GiB8,782,409,0882.561—bartowski
IQ2_M8.84 GiB9,493,073,2802.768—bartowski
UD-IQ2_M8.96 GiB9,624,505,4402.807—unsloth
Q2_K9.78 GiB10,503,720,6723.063—unsloth
Q2_K_L9.78 GiB10,503,720,6723.063—unsloth
Q2_K9.78 GiB10,503,720,9603.063—bartowski
IQ3_XXS9.98 GiB10,716,478,8483.125—bartowski
UD-IQ3_XXS10.07 GiB10,810,020,9603.152—unsloth
Q2_K_L10.10 GiB10,845,115,7763.163—bartowski
IQ3_XS10.77 GiB11,562,233,8563.372—bartowski
Q3_K_S11.33 GiB12,167,614,1763.548—unsloth
Q3_K_S11.33 GiB12,167,614,4643.548—bartowski
IQ3_M11.69 GiB12,547,074,0483.659—bartowski
Q3_K_M12.51 GiB13,437,640,4163.919—unsloth
Q3_K_M12.51 GiB13,437,640,7043.919—bartowski
Q3_K_L13.54 GiB14,543,462,4004.241—bartowski
IQ4_XS13.75 GiB14,767,447,7764.307—unsloth
IQ4_XS13.75 GiB14,767,448,0644.307—bartowski
IQ4_NL14.50 GiB15,567,396,5764.540—unsloth
IQ4_NL14.50 GiB15,567,396,8644.540—bartowski
Q4_014.55 GiB15,617,973,9844.555—unsloth
Q4_014.55 GiB15,617,974,2724.555—bartowski
Q4_K_S14.60 GiB15,674,056,4164.571—unsloth
Q4_K_S14.60 GiB15,674,056,7044.571—bartowski
Q4_K_M15.41 GiB16,546,688,7364.825—unsloth
Q4_K_M15.41 GiB16,546,689,0244.825—bartowski
Q4_K_L15.73 GiB16,888,083,8404.925—bartowski
Q4_115.99 GiB17,167,294,1765.006—unsloth
Q4_115.99 GiB17,167,294,4645.006—bartowski
Q5_K_S17.48 GiB18,767,191,7765.473—unsloth
Q5_K_S17.48 GiB18,767,192,0645.473—bartowski
Q5_K_M17.95 GiB19,271,675,6165.620—unsloth
Q5_K_M17.95 GiB19,271,675,9045.620—bartowski
Q5_K_L18.27 GiB19,613,070,7205.720—bartowski
Q6_K20.64 GiB22,166,974,1766.465—unsloth

Compare with

same modality, comparable size
vs Qwen3.6-35B-A3Bvs Qwen3.6-27B-Fable-Fusion-711-Uncensored-Heretic-NM-DAU-MTPvs Qwen3.5-9Bvs gemma-4-26B-A4B-itvs gemma-4-12B-itvs Qwen3.5-4BAll alternatives →Will it fit on my card? →

Will it run on your card?

full quant x context sweep
By VRAM:8GB12GB16GB24GB32GB48GB
Radeon RX 6500 XT 4GBGeForce RTX 3050 6GBGeForce RTX 5050 8GBGeForce RTX 3080 10GBGeForce RTX 2080 Ti 11GBGeForce RTX 5070 12GBGeForce RTX 5060 Ti 16GBApple M3 Pro 18GBGeForce RTX 3080 Ti 20GBGeForce RTX 5090 D V2 24GBGeForce RTX 5090 D 32GBApple M5 Max 36GBApple M5 Pro 48GBApple M5 Max 64GBApple M3 Ultra 96GBApple M5 Max 128GBApple M2 Ultra 192GBApple M3 Ultra 256GBApple M3 Ultra 512GB

Why other calculators give a different number

A parameters × bits ÷ 8 estimate puts Q4_K_M at roughly 14.37 GiB. The real file is 15.41 GiB, because a quantization is a mixture and some tensors are always kept at higher precision.

Architecture

Architecture unavailable — this repository is gated and no ungated mirror was found. Exact file sizes above are still authoritative; only the KV math needs the config.

Questions people ask

How much VRAM does gemma-3-27b-it-qat-q4_0-unquantized need?
Q4_K_M is exactly 16,546,688,736 bytes (15.41 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
Which quantization of gemma-3-27b-it-qat-q4_0-unquantized should I use?
Q4_K_M is the usual default. Pick the largest quantization that fits your card at the context you actually need — the table above gives exact sizes for every one published.