nbeerbower · text

Gemma4-Gutenberg-26B-A4B-lora

nbeerbower/Gemma4-Gutenberg-26B-A4B-lora

Gemma4-Gutenberg-26B-A4B-lora at Q4_K_M is exactly 16,796,016,000 bytes (15.64 GiB / 16.80 GB) — an effective 5.325 bits per weight, not the nominal 4.

From the file· summed from 1 file(s)
Parameters
25.2B
Architecture
gemma4
Context
—
native (config.json)
License
gemma

Shipped quantizations

● exact bytes, summed from published files
QuantSize●Exact bytes●Effective bpw●Tensors●Publisher
I1-IQ1_S7.72 GiB8,290,271,3922.628—spiritfather
I1-IQ1_M8.07 GiB8,668,657,3122.748—spiritfather
I1-IQ2_XXS8.66 GiB9,299,300,5122.948—spiritfather
I1-IQ2_XS9.14 GiB9,816,430,7523.112—spiritfather
I1-IQ2_S9.20 GiB9,873,201,3123.130—spiritfather
I1-IQ2_M9.67 GiB10,377,715,8723.290—spiritfather
Q2_K9.86 GiB10,582,737,2803.355—spiritfather
I1-Q2_K9.86 GiB10,582,737,5683.355—spiritfather
I1-Q2_K_S9.89 GiB10,624,481,9523.368—spiritfather
I1-IQ3_XXS10.55 GiB11,325,694,1123.591—spiritfather
I1-IQ3_XS10.84 GiB11,636,068,0003.689—spiritfather
Q3_K_S11.38 GiB12,222,409,6003.875—spiritfather
I1-Q3_K_S11.38 GiB12,222,409,8883.875—spiritfather
I1-IQ3_S11.38 GiB12,222,409,8883.875—spiritfather
I1-IQ3_M11.54 GiB12,392,563,8723.929—spiritfather
Q3_K_M12.37 GiB13,286,733,6964.213—spiritfather
I1-Q3_K_M12.37 GiB13,286,733,9844.213—spiritfather
Q3_K_L12.88 GiB13,824,488,3204.383—spiritfather
I1-Q3_K_L12.88 GiB13,824,488,6084.383—spiritfather
I1-IQ4_XS12.96 GiB13,917,726,3684.412—spiritfather
IQ4_XS13.10 GiB14,063,808,8964.459—spiritfather
I1-Q4_013.49 GiB14,488,056,4804.593—spiritfather
Q4_K_S14.40 GiB15,464,825,2164.903—spiritfather
I1-Q4_K_S14.40 GiB15,464,825,5044.903—spiritfather
I1-Q4_114.87 GiB15,969,576,6085.063—spiritfather
Q4_K_M15.64 GiB16,796,016,0005.325—spiritfather
I1-Q4_K_M15.64 GiB16,796,016,2885.325—spiritfather
Q5_K_S16.75 GiB17,986,733,4405.703—spiritfather
I1-Q5_K_S16.75 GiB17,986,733,7285.703—spiritfather
Q5_K_M17.82 GiB19,132,890,4966.066—spiritfather
I1-Q5_K_M17.82 GiB19,132,890,7846.066—spiritfather
Q6_K21.08 GiB22,638,399,3607.177—spiritfather
I1-Q6_K21.08 GiB22,638,399,6487.177—spiritfather
Q8_025.02 GiB26,859,859,3288.516—spiritfather

Compare with

same modality, comparable size
vs Qwen3-Coder-30B-A3B-Instructvs Qwen3.6-27Bvs Qwen3.8-27Bvs gemma-4-12B-it-qat-q4_0-unquantizedvs gemma-4-E4B-itvs Qwen3-30B-A3B-Thinking-2507All 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 13.22 GiB. The real file is 15.64 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 Gemma4-Gutenberg-26B-A4B-lora need?
Q4_K_M is exactly 16,796,016,000 bytes (15.64 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
Which quantization of Gemma4-Gutenberg-26B-A4B-lora 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.