What GPU do I need to run cyankiwi/gemma-4-E4B-it-AWQ-INT4?

8.2B parameters, published in F16. View on Hugging Face

8.2B
Parameters
F16
Native precision
Gemma4ForConditionalGeneration
Architecture
any-to-any
Pipeline

gemma-4-E4B-it-AWQ-INT4 is published by cyankiwi on Hugging Face, with 185,105 downloads and 4 likes to date. It's a Gemma4ForConditionalGeneration model built for any-to-any, published natively in F16. Its name carries the AWQ tag.

VRAM required & cheapest live GPU fit

Required VRAM = weight size at each precision times a flat 1.2 overhead for activations and fragmentation. The KV-cache grows with context and is not in that factor; it is listed per model below where the architecture is published. Full formula and assumptions: methodology.

PrecisionWeight sizeRequired VRAMCheapest live fitGPUs neededEst. $/hr (full fit)
FP1615.2 GB18.3 GBRTX A50001$0.176/hr
FP8 (quantized)7.6 GB9.1 GBRTX 5060 Ti1$0.180/hr
INT4 (quantized)3.8 GB4.6 GBRTX 30701$0.088/hr

A GPU is only matched to a row if its hardware supports that precision, and the primary recommendation is always a single-GPU fit when one exists.

INT4 requires a quantized checkpoint actually published for this model, check its Hugging Face page before relying on this row.

Cheapest way to run gemma-4-E4B-it-AWQ-INT4 at its published (F16) precision: 1× RTX A5000, at $0.176/hr per GPU ($0.176/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.

gemma-4-E4B-it-AWQ-INT4: KV cache by context length

The KV cache is the memory the attention layers hold for every token of context, on top of the weights. It is not part of the flat 1.2x overhead in the table above, grows with context length and with every concurrent request, and is why a long-context deployment needs more VRAM than the table shows.

Sliding-window attention: windowed layers stop growing at the window, the full-attention layers keep growing. Cached per token: 86,016 bytes at 16-bit while the context is inside the window.

ContextKV cache, one sequence
4K0.09 GB
32K0.47 GB
128K1.8 GB

Computed from the layer, head and window counts in the model's own configuration (cyankiwi/gemma-4-E4B-it-AWQ-INT4), at 16 bits per cached value. Each concurrent sequence needs its own cache, on top of the weights.

gemma-4-E4B-it-AWQ-INT4: common questions

Does gemma-4-E4B-it-AWQ-INT4 fit on a 24 GB GPU?

Yes. At FP16 it needs 18.3 GB of VRAM, so a 24 GB card holds it with 5.7 GB to spare. A 16 GB card is not enough for it at FP16.

What is the least VRAM gemma-4-E4B-it-AWQ-INT4 can run in?

4.6 GB, at INT4 (quantized), which fits a 6 GB card, against 18.3 GB at FP16. That row assumes an INT4 checkpoint has actually been published for this model, so check its Hugging Face page before planning around it.

Does quantizing gemma-4-E4B-it-AWQ-INT4 lower the GPU bill?

Yes. At FP16 the cheapest live fit is one RTX A5000 at $0.176/hr. At INT4 (quantized) it drops to one RTX 3070 at $0.088/hr, provided a quantized checkpoint exists for it.

Weight-to-VRAM math, the fit rules, and how live prices are normalized: full methodology.

More Gemma 4 models

All 44 Gemma 4 models: VRAM and GPU requirements

Lineage

Based on gemma-4-E4B-it.

Alternatives at this size

Other models for any-to-any within about a third of gemma-4-E4B-it-AWQ-INT4's 8.2B parameters, from other model lines.

More on gemma-4-E4B-it-AWQ-INT4

Related reading: RTX A5000 pricing and specs, How much VRAM you need for LLMs, Serving LLMs with vLLM, vLLM vs TensorRT-LLM vs SGLang, and Best GPU for LLM inference.

Your own GPU cloud accounts

One console for the GPU clouds you already rent from. Your keys stay on your machine.

$20 a month, flat

Connect your accounts

One console for the GPU clouds you already use.

Connect your accounts for $20 a month, flat. Your provider bills the GPU.

© 2026 Aquanode. All rights reserved.

All trademarks, logos and brand names are the property of their respective owners.