google · text

datagemma-rag-27b-it

google/datagemma-rag-27b-it

datagemma-rag-27b-it at Q4_K_M is exactly 16,645,382,208 bytes (15.50 GiB / 16.65 GB) — an effective 4.891 bits per weight, not the nominal 4.

From the file· summed from 1 file(s)
Parameters
27.2B
Architecture
gemma2
Context
—
native (config.json)
License
gemma

Shipped quantizations

● exact bytes, summed from published files
QuantSize●Exact bytes●Effective bpw●Tensors●Publisher
IQ1_S5.71 GiB6,132,432,9601.802—legraphista
IQ1_M6.23 GiB6,692,470,8481.966—legraphista
IQ2_XXS7.10 GiB7,625,867,3282.241—legraphista
IQ2_XXS7.10 GiB7,625,867,3282.241—bartowski
IQ2_XS7.82 GiB8,399,716,4162.468—legraphista
IQ2_S8.06 GiB8,652,161,0882.542—legraphista
IQ2_M8.75 GiB9,398,878,2722.762—bartowski
IQ2_M8.75 GiB9,398,878,2722.762—legraphista
Q2_K_S9.06 GiB9,722,765,3762.857—legraphista
Q2_K9.73 GiB10,449,576,0003.070—legraphista
Q2_K9.73 GiB10,449,576,0003.070—bartowski
Q2_K_L10.00 GiB10,735,272,0003.154—bartowski
IQ3_XXS10.01 GiB10,750,754,8803.159—legraphista
IQ3_XS10.76 GiB11,550,629,9523.394—legraphista
IQ3_XS10.76 GiB11,550,629,9523.394—bartowski
IQ3_S11.33 GiB12,169,060,4163.576—legraphista
Q3_K_S11.33 GiB12,169,060,4163.576—bartowski
Q3_K_S11.33 GiB12,169,060,4163.576—legraphista
IQ3_M11.60 GiB12,454,830,1443.659—bartowski
IQ3_M11.60 GiB12,454,830,1443.659—legraphista
Q3_K_M12.50 GiB13,424,648,2563.945—bartowski
Q3_K12.50 GiB13,424,648,2563.945—legraphista
Q3_K_L13.52 GiB14,519,361,6004.266—bartowski
Q3_K_L13.52 GiB14,519,361,6004.266—legraphista
IQ4_XS13.80 GiB14,814,421,0564.353—legraphista
IQ4_XS13.80 GiB14,814,421,0564.353—bartowski
IQ4_NL14.56 GiB15,628,378,1764.592—legraphista
Q4_014.60 GiB15,681,462,3364.608—bartowski
Q4_K_S14.66 GiB15,739,265,0884.625—bartowski
Q4_K_S14.66 GiB15,739,265,0884.625—legraphista
Q4_K15.50 GiB16,645,382,2084.891—legraphista
Q4_K_M15.50 GiB16,645,382,2084.891—bartowski
Q4_K_L15.77 GiB16,931,078,2084.975—bartowski
Q5_K_S17.59 GiB18,884,206,4005.549—legraphista
Q5_K_S17.59 GiB18,884,206,6565.549—bartowski
Q5_K18.08 GiB19,408,117,5685.703—legraphista
Q5_K_M18.08 GiB19,408,117,8245.703—bartowski
Q5_K_L18.34 GiB19,693,813,8245.787—bartowski
Q6_K20.81 GiB22,343,524,1606.565—legraphista
Q6_K20.81 GiB22,343,524,4166.565—bartowski

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 14.26 GiB. The real file is 15.50 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 datagemma-rag-27b-it need?
Q4_K_M is exactly 16,645,382,208 bytes (15.50 GiB) in weights. Add the KV cache, which depends on your context length, plus roughly half a gigabyte of runtime overhead.
Which quantization of datagemma-rag-27b-it 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.