What GPU do I need to run EleutherAI/gpt-neox-20b?

20.7B parameters, published in F16. View on Hugging Face

20.7B
Parameters
F16
Native precision
GPTNeoXForCausalLM
Architecture
text-generation
Pipeline

gpt-neox-20b is published by EleutherAI on Hugging Face, with 787,110 downloads and 585 likes to date. It's a GPTNeoXForCausalLM model built for text-generation, published natively in F16.

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)
FP1638.6 GB46.4 GBRTX A60001$0.363/hr
cheaper alt.V1003$0.330/hr
FP8 (quantized)19.3 GB23.2 GBRTX 40901$0.374/hr
INT4 (quantized)9.7 GB11.6 GBRTX A50001$0.176/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 gpt-neox-20b at its published (F16) precision: 1× RTX A6000, at $0.363/hr per GPU ($0.363/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.

gpt-neox-20b: 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.

Multi-head attention: every layer caches keys and values for every head. Cached per token: 1,081,344 bytes at 16-bit.

ContextKV cache, one sequence
2K (model maximum)2.1 GB

Computed from the layer, head and window counts in the model's own configuration (EleutherAI/gpt-neox-20b), at 16 bits per cached value. Each concurrent sequence needs its own cache, on top of the weights.

gpt-neox-20b: common questions

Can gpt-neox-20b run on a single GPU?

Yes, but not on a desktop card. At FP16 it needs 46.4 GB of VRAM, more than a single 32 GB desktop card holds. The cheapest live single-card fit is a 48.0 GB RTX A6000 at $0.363/hr.

What is the least VRAM gpt-neox-20b can run in?

11.6 GB, at INT4 (quantized), which fits a 12 GB card, against 46.4 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 gpt-neox-20b lower the GPU bill?

Yes. At FP16 the cheapest live fit is one RTX A6000 at $0.363/hr. At INT4 (quantized) it drops to one RTX A5000 at $0.176/hr, provided a quantized checkpoint exists for it.

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

More GPT-NeoX models

All 2 GPT-NeoX models: VRAM and GPU requirements

Alternatives at this size

Other models for text-generation within about a third of gpt-neox-20b's 20.7B parameters, from other model lines.

More on gpt-neox-20b

Related reading: RTX A6000 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.