What GPU do I need to run fishaudio/drama-3?

35.9B parameters, published in F8_E4M3. View on Hugging Face

35.9B
Parameters
F8_E4M3
Native precision
FishOmniModel
Architecture
text-to-speech
Pipeline

drama-3 is published by fishaudio on Hugging Face, with 44,913 downloads and 1,228 likes to date. It's a FishOmniModel model built for text-to-speech, published natively in F8_E4M3.

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)
FP8 (native)33.4 GB40.1 GBL401$0.759/hr
cheaper alt.RTX 5060 Ti3$0.541/hr
INT4 (quantized)16.7 GB20.1 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 drama-3 at its published (F8_E4M3) precision: 1× L40, at $0.759/hr per GPU ($0.759/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.

drama-3: common questions

Can drama-3 run on a single GPU?

Yes, but not on a desktop card. At FP8 (native) it needs 40.1 GB of VRAM, more than a single 32 GB desktop card holds. The cheapest live single-card fit is a 48.0 GB L40 at $0.759/hr.

Is drama-3 already quantized?

Yes. It is published in FP8, one byte per parameter, so the 40.1 GB figure is already a quantized footprint rather than a full-precision one. Only INT4 goes below it, at 20.1 GB. A GPU without FP8 tensor cores cannot run it as published, which is why cards here are matched on precision support and not on VRAM alone.

What is the least VRAM drama-3 can run in?

20.1 GB, at INT4 (quantized), which fits a 24 GB card, against 40.1 GB at FP8 (native). 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 drama-3 lower the GPU bill?

Yes. At FP8 (native) the cheapest live fit is one L40 at $0.759/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 fishaudio models

All 2 fishaudio models: VRAM and GPU requirements

More on drama-3

Related reading: L40 pricing and specs, Whisper variants compared, and What AI inference is.

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.