SANA 1.5 4.8B TPS calculator

Open weights NVIDIA,Massachusetts Institute of Technology (MIT),Tsinghua University,Playground,Peking University,The University of Hong Kong 4.8B parameters May 2025

Each card below is assessed against this model at the context length and minimum quality you choose. Speed is an estimate for a single request, calculated from the card's memory bandwidth and the size of the model once compressed.

Calculated for this model

818 cards that can run it

818 cards we hold specifications for

Smallest card that fits

Tesla C1080

4 GB · IQ4_XS · 18.9 tok/s

Fastest card

B200

706 tok/s · 180 GB

Which GPUs can run SANA 1.5 4.8B?

Set the inputs, read the answer

Context length tokens

A longer conversation needs more memory, which can push this model off smaller cards.

Minimum quality

Hides cards that would only fit the model by compressing it below this point.

Search cards

Card memory

Also show cards that cannot run it

818 cards match

Calculating

					Needs	Quantisation	Fit
706 tok/s 424–1,129 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	5.8 GB	Q8_0	Comfortable
706 tok/s 424–1,129 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	5.8 GB	Q8_0	Comfortable
564 tok/s 338–902 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	5.8 GB	Q8_0	Comfortable
564 tok/s 338–902 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	5.8 GB	Q8_0	Comfortable
451 tok/s 270–721 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	5.8 GB	Q8_0	Comfortable
431 tok/s 259–690 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	5.8 GB	Q8_0	Comfortable
431 tok/s 259–690 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	5.8 GB	Q8_0	Comfortable
413 tok/s 248–661 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	5.8 GB	Q8_0	Comfortable
366 tok/s 220–586 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	5.8 GB	Q8_0	Comfortable
366 tok/s 220–586 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	5.8 GB	Q8_0	Comfortable
366 tok/s 220–586 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	5.8 GB	Q8_0	Comfortable
348 tok/s 209–556 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	5.8 GB	Q8_0	Comfortable
296 tok/s 178–474 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	5.8 GB	Q8_0	Comfortable
296 tok/s 178–474 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	5.8 GB	Q8_0	Comfortable
296 tok/s 178–474 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	5.8 GB	Q8_0	Comfortable
296 tok/s 178–474 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	5.8 GB	Q8_0	Comfortable
296 tok/s 178–474 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	5.8 GB	Q8_0	Comfortable
226 tok/s 135–361 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	5.8 GB	Q8_0	Comfortable
226 tok/s 135–361 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	5.8 GB	Q8_0	Comfortable
188 tok/s 113–301 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	5.8 GB	Q8_0	Comfortable
184 tok/s 110–295 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	5.8 GB	Q8_0	Comfortable
180 tok/s 108–288 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	5.8 GB	Q8_0	Comfortable
180 tok/s 108–288 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	5.8 GB	Q8_0	Comfortable
180 tok/s 108–288 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	5.8 GB	Q8_0	Comfortable
180 tok/s 108–288 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	5.8 GB	Q8_0	Comfortable

Speeds are estimates for a single request — one conversation at a time — calculated from memory bandwidth, model size and quantisation. Real throughput varies with the inference runtime and its version. Figures published by hardware vendors measure many simultaneous requests and are much higher.

On record

Full specification

Everything on record for this model. Most of it describes how it was trained rather than how it runs — useful context for judging how much work went into it, and how it compares with models built at a different scale.

Origin

Who built this model, where, and when it was published.

Organisation: NVIDIA,Massachusetts Institute of Technology (MIT),Tsinghua University,Playground,Peking University,The University of Hong Kong
Organisation type: Industry,Academia,Academia,Industry,Academia,Academia
Country: United States of America, China, Hong Kong
Published: 17 May 2025
Authors: Enze Xie, Junsong Chen, Yuyang Zhao, Jincheng Yu, Ligeng Zhu, Chengyue Wu, Yujun Lin, Zhekai Zhang, Muyang Li, Junyu Chen, Han Cai, Bingchen Liu, Daquan Zhou, Song Han

What it does

The problem areas the model was built for. A model can carry several of each.

Domain: Image generation
Task: Text-to-image, Image generation
Base model: SANA 1.6B
Numerical format: BF16

Size

How large the model is and how much data it was trained on. Parameters are the figure that decides whether it fits on a given graphics card.

Parameters: 4.8B
Training data: tokens

The training run

What it physically took to train: which chips, how many, for how long, and what that drew from the wall.

Training hardware: NVIDIA A100
Chips used: 64
Power draw: 50.2 kW

Availability

Whether you can obtain the model and run it on your own hardware, which is what decides if any of the graphics-card figures on this page apply.

Weights: Open — downloadable
Model access: Open weights (non-commercial)
Training code: Open source
Hugging Face: Efficient-Large-Model

How it is classified

Labels the source dataset applies when tracking notable models, and how confident it is in the entry.

Record confidence: Confident

Sources

Where this record came from and when it was last checked.

Reference: SANA 1.5: Efficient Scaling of Training-Time and Inference-Time Compute in Linear Diffusion Transformer
Last updated: 28 November 2025

The extremes

The ten fastest GPUs that run SANA 1.5 4.8B

Ranked by estimated tokens per second, newest card first where speeds tie. Because generation is bound by memory bandwidth, this ordering follows bandwidth rather than any gaming benchmark.

The smallest GPUs that still run SANA 1.5 4.8B

The cheapest route in, by memory capacity. A tight fit runs the model but leaves nothing spare for a longer conversation.

What the numbers mean

Hardware requirements in practice

Minimum card

Tesla C1080

Memory needed

3.3 GB

Fastest

706 tok/s

SANA 1.5 4.8B reaches a parameter count of 4.8B. That is small enough that hardware is rarely the obstacle, including on cards several years old. The number of cards we track that can run it: 818.

The entry point is Tesla C1080, with a memory capacity of 4 GB, running it at a compression of IQ4_XS and producing around 18.9 tokens per second.

The quickest result comes from B200, generating roughly 706 tokens per second on the strength of a memory bandwidth of 8,000 GB/s.

Background

SANA 1.5 4.8B was published by NVIDIA,Massachusetts Institute of Technology (MIT),Tsinghua University,Playground,Peking University,The University of Hong Kong, in the country recorded as United States of America, during May 2025. The publishing organisation is categorised as industry,Academia,Academia,Industry,Academia,Academia.

It works in the domain of Image generation, and is recorded as performing the task of text-to-image, Image generation.

It builds on SANA 1.6B. Most models at this scale are adapted from an existing base rather than built from nothing.

Because its weights were released, nothing about running it depends on a provider staying available — it is yours once downloaded. On Hugging Face it is published under the organisation Efficient-Large-Model.

Reading the throughput figures

Across every card that can run it, the middle of the range sits at 28.2 tokens per second. Producing text faster than most people read it: 771 of them.

Every weight participates in every token here, so bandwidth is the whole story: the ranking below is effectively a ranking of memory throughput.

Its internal architecture is not on file, so memory is approximated from the parameter count and marked accordingly. Expect the real figure to differ, more so at long context.

Step by step

How to choose a GPU for SANA 1.5 4.8B

The table above has already assessed every card we hold specifications for against this model. Getting to your answer takes six steps.

01

Read the memory figure first

Every card here has been checked against SANA 1.5 4.8B, needing around 3.3 GB at a compression of IQ4_XS. That figure, not the headline performance of a card, is what decides whether it runs.
02

Set the context length you will work at

Longer conversations cost memory on top of what the weights need. Move the slider to your real working length before trusting any row for SANA 1.5 4.8B.
03

Decide how much compression you will accept

Compression is what makes a model fit smaller cards, at some cost in accuracy, reaching a compression of IQ4_XS on the smallest card that fits. Setting a minimum quality drops the cards that only manage it by squeezing further than you would want, and holds the comparison at one level.
04

Rank by throughput rather than spec sheet

Ranking by tokens per second follows memory bandwidth rather than core counts, for SANA 1.5 4.8B. It will not match a gaming ordering, because generation is bound by memory bandwidth. The card topping the list is B200, at 706 tok/s.
05

Look at the headroom, not just the fit

Tight means it loads and works with no room to raise the context later, in the case of SANA 1.5 4.8B. Comfortable means you can grow the context later. That difference matters more than a few tokens per second, so buy for comfortable if you expect to.
06

See what else that card runs

Following a card through to its own page shows every other model it can hold, which is the question that follows once you have settled on SANA 1.5 4.8B.

Answers

SANA 1.5 4.8B — common questions

SANA 1.5 4.8B— can I run it if it does not fit in my GPU?

It can be split between the card and system memory, but it generates painfully slowly that way. Every figure here assumes the whole model is resident on the card.

SANA 1.5 4.8B— would two GPUs run it faster?

A second card roughly doubles the memory available but not the generation rate. The number already able to run it alone: 818. So a second card is rarely the answer here.

SANA 1.5 4.8B— why does the quantisation differ between cards?

Each card is shown running the least-compressed copy it can hold. The number of distinct compression levels across the cards that fit it: 4. Bigger cards get the more accurate version, and the quality floor above pins the comparison to one level.

SANA 1.5 4.8B— how accurate are these speed estimates?

They are calculated from specifications rather than measured, and each carries a range. One example: 424–1,129 tok/s on B200. The same model and card vary by thirty to fifty per cent depending on the inference software and its version.

SANA 1.5 4.8B— what GPU do I need to run it?

The smallest card in our catalogue that holds it is Tesla C1080, with a memory capacity of 4 GB. It runs the model at a compression of IQ4_XS using about 3.3 GB, and produces roughly 18.9 tokens per second. The number of cards able to run it in total: 818.

SANA 1.5 4.8B— how fast is it on a GPU?

It depends on the card. The quickest we calculate is B200, at about 706 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and the number of cards clearing that: 771.

SANA 1.5 4.8B— how much VRAM does it need?

It needs about 3.3 GB at a compression of IQ4_XS, which is what the smallest card that runs it uses. Less compression needs more: the figures in the memory column above are recalculated for each card, because each one holds the least-compressed version it can.

SANA 1.5 4.8B— can I run it on a GPU holding 8 GB?

Yes. The card CMP 170HX 8 GB, holding 8 GB, runs it at a compression of Q8_0, using about 5.8 GB and generating roughly 131 tokens per second. The fit is comfortable.

SANA 1.5 4.8B— can I run it on a GPU holding 12 GB?

Yes. The card GeForce RTX 3080 Ti, holding 12 GB, runs it at a compression of Q8_0, using about 5.8 GB and generating roughly 80.5 tokens per second. The fit is comfortable.

SANA 1.5 4.8B— can I run it on a GPU holding 16 GB?

Yes. The card Tesla V100 SXM2 16 GB, holding 16 GB, runs it at a compression of Q8_0, using about 5.8 GB and generating roughly 99.7 tokens per second. The fit is comfortable.

SANA 1.5 4.8B— can I run it on a GPU holding 24 GB?

Yes. The card GeForce RTX 5090 D V2, holding 24 GB, runs it at a compression of Q8_0, using about 5.8 GB and generating roughly 118 tokens per second. The fit is comfortable.

SANA 1.5 4.8B— is it open source?

Its weights are published, so it can be downloaded and run on your own hardware. Note that open weights is not the same as open source in the full sense — it says nothing about the training data, the training code, or the commercial terms attached.

SANA 1.5 4.8B— how many parameters does it have?

It has a parameter count of 4.8B. 4.8B Our final model (SANA-4.8B) scales to 60 layers while maintaining the same channel dimension (2240 per layer) and FFN dimension (5600) as SANA-1.6B. The architecture, training data, and other hyperparameters remain consistent with SANA-1.6B. That figure is the total, and it is what decides how much memory the model needs — roughly half a gigabyte per billion at the compression most people use.

SANA 1.5 4.8B— who created it?

It was published by NVIDIA,Massachusetts Institute of Technology (MIT),Tsinghua University,Playground,Peking University,The University of Hong Kong, based in United States of America, an organisation categorised as industry,Academia,Academia,Industry,Academia,Academia.

SANA 1.5 4.8B— when was it released?

It was published in May 2025.

SANA 1.5 4.8B— what is it used for?

It works in the domain of Image generation, and is recorded as handling the task of text-to-image, Image generation. A model can carry several of each, so these are the areas it was built for rather than a limit on what it will attempt.

SANA 1.5 4.8B— where can I download it?

Its weights are published on Hugging Face, under the organisation Efficient-Large-Model. We do not host model files — this site calculates what hardware is needed to run them.

Source

Original publication

Record last updated 28 November 2025

The other direction

Looking at it from the other side?

This page starts from the model. If you already own a card and want to know everything it will run, start from the hardware instead.

All AI models