What GPU do I need to run google-t5/t5-3b?
2.9B parameters, published in F32. View on Hugging Face
t5-3b is published by google-t5 on Hugging Face, with 493,115 downloads and 52 likes to date. It's a T5WithLMHeadModel model built for translation, published natively in F32.
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) |
|---|---|---|---|---|---|
| FP32 | 10.6 GB | 12.7 GB | V100 | 1 | $0.110/hr |
| FP8 (quantized) | 2.7 GB | 3.2 GB | RTX 5060 Ti | 1 | $0.145/hr |
| INT4 (quantized) | 1.3 GB | 1.6 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 t5-3b at its published (F32) precision: 1× V100, at $0.110/hr per GPU ($0.110/hr total). Quantizing to FP8 or INT4 (rows above) can cost less, but requires a compatible quantized checkpoint to exist for this model.
t5-3b: common questions
Does t5-3b fit on a 16 GB GPU?
Yes. At FP32 it needs 12.7 GB of VRAM, so a 16 GB card holds it with 3.3 GB to spare. A 12 GB card is not enough for it at FP32.
Can t5-3b run in 16-bit instead of FP32?
Yes. Its published weights are FP32, 10.6 GB, or 12.7 GB once inference overhead is added. Loading the same weights in 16 bits halves that to 5.3 GB, or 6.4 GB with overhead. That moves it onto an 8 GB card instead of a 16 GB one. How much accuracy the cast costs is model-specific and is not measured here.
What is the least VRAM t5-3b can run in?
1.6 GB, at INT4 (quantized), which fits a 6 GB card, against 12.7 GB at FP32. 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 t5-3b lower the GPU bill?
Yes. At FP32 the cheapest live fit is one V100 at $0.110/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 T5 models
- T5_Paraphrase_Paws (223M, F32)
- chatgpt_paraphraser_on_T5_base (223M, F32)
Alternatives at this size
Other models for translation within about a third of t5-3b's 2.9B parameters, from other model lines.
- Hy-MT2-1.8B (2.0B, BF16)
- Index-Translate-2B (2.3B, BF16)
More on t5-3b
Fits on a 16 GB GPU at FP32: every model that fits in 16 GB.
Fits on an 8 GB GPU at FP8: every model that fits in 8 GB.
Best translation models: how t5-3b ranks against the rest.
Related reading: V100 pricing and specs, How much VRAM you need for LLMs, 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