What GPU do I need to run higurashi-ai/Haru-4B?
5.3B parameters, published in BF16. View on Hugging Face
Haru-4B is published by higurashi-ai on Hugging Face, with 2,098 downloads and 329 likes to date. It's a HaruForConditionalGeneration model built for audio-to-audio, published natively in BF16.
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.
| Precision | Weight size | Required VRAM | Cheapest live fit | GPUs needed | Est. $/hr (full fit) |
|---|---|---|---|---|---|
| BF16 | 9.8 GB | 11.8 GB | RTX A4000 | 1 | $0.132/hr |
| FP8 (quantized) | 4.9 GB | 5.9 GB | RTX 5060 Ti | 1 | $0.180/hr |
| INT4 (quantized) | 2.5 GB | 2.9 GB | RTX 3070 | 1 | $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 Haru-4B at its published (BF16) precision: 1× RTX A4000, at $0.132/hr per GPU ($0.132/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.
Haru-4B: common questions
Does Haru-4B fit on a 12 GB GPU?
Yes. At BF16 it needs 11.8 GB of VRAM, so a 12 GB card holds it with 0.2 GB to spare. An 8 GB card is not enough for it at BF16.
What is the least VRAM Haru-4B can run in?
2.9 GB, at INT4 (quantized), which fits a 6 GB card, against 11.8 GB at BF16. 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 Haru-4B lower the GPU bill?
Yes. At BF16 the cheapest live fit is one RTX A4000 at $0.132/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 on Haru-4B
Fits on a 12 GB GPU at BF16: every model that fits in 12 GB.
Fits on an 8 GB GPU at FP8: every model that fits in 8 GB.
Related reading: RTX A4000 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