What GPU do I need to run BreezeBlue/Breeze-TTS-2?

3.5B parameters, published in BF16. View on Hugging Face

3.5B
Parameters
BF16
Native precision
BreezeForConditionalGeneration
Architecture
text-to-speech
Pipeline

Breeze-TTS-2 is published by BreezeBlue on Hugging Face, with 3,086 downloads and 328 likes to date. It's a BreezeForConditionalGeneration model built for text-to-speech, 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.

PrecisionWeight sizeRequired VRAMCheapest live fitGPUs neededEst. $/hr (full fit)
BF166.5 GB7.7 GBRTX 30701$0.088/hr
FP8 (quantized)3.2 GB3.9 GBRTX 5060 Ti1$0.145/hr
INT4 (quantized)1.6 GB1.9 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 Breeze-TTS-2 at its published (BF16) precision: 1× RTX 3070, at $0.088/hr per GPU ($0.088/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.

Breeze-TTS-2: 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.

Grouped-query attention: every layer caches keys and values for a few shared KV heads. Cached per token: 114,688 bytes at 16-bit.

ContextKV cache, one sequence
2K (model maximum)0.22 GB

Computed from the layer, head and window counts in the model's own configuration (BreezeBlue/Breeze-TTS-2), at 16 bits per cached value. Each concurrent sequence needs its own cache, on top of the weights.

Breeze-TTS-2: common questions

Does Breeze-TTS-2 fit on a 8 GB GPU?

Yes. At BF16 it needs 7.7 GB of VRAM, so an 8 GB card holds it with 0.3 GB to spare. A 6 GB card is not enough for it at BF16.

What is the least VRAM Breeze-TTS-2 can run in?

1.9 GB, at INT4 (quantized), which fits a 6 GB card, against 7.7 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.

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

BreezeBlue models

Alternatives at this size

Other models for text-to-speech within about a third of Breeze-TTS-2's 3.5B parameters, from other model lines.

More on Breeze-TTS-2

Related reading: RTX 3070 pricing and specs, Whisper variants compared, and What AI inference is.

Related on Aquanode: All bosonai models and their GPU requirements.

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.