SimPO TPS calculator

Open weights Princeton University,University of Virginia 9B parameters November 2024

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

582 of 818 cards that can run it

Smallest card that fits

Quadro 6000

6 GB · Q3_K_M · 15.5 tok/s

Fastest card

B200

376 tok/s · 180 GB

Which GPUs can run SimPO?

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

582 cards match

Calculating

					Needs	Quantisation	Fit
376 tok/s 226–602 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	10.3 GB	Q8_0	Comfortable
376 tok/s 226–602 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	10.3 GB	Q8_0	Comfortable
301 tok/s 180–481 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	10.3 GB	Q8_0	Comfortable
301 tok/s 180–481 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	10.3 GB	Q8_0	Comfortable
240 tok/s 144–385 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	10.3 GB	Q8_0	Comfortable
230 tok/s 138–368 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	10.3 GB	Q8_0	Comfortable
230 tok/s 138–368 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	10.3 GB	Q8_0	Comfortable
220 tok/s 132–352 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	10.3 GB	Q8_0	Comfortable
195 tok/s 117–313 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	10.3 GB	Q8_0	Comfortable
195 tok/s 117–313 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	10.3 GB	Q8_0	Comfortable
195 tok/s 117–313 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	10.3 GB	Q8_0	Comfortable
185 tok/s 111–297 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	10.3 GB	Q8_0	Comfortable
158 tok/s 95–253 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	10.3 GB	Q8_0	Comfortable
158 tok/s 95–253 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	10.3 GB	Q8_0	Comfortable
158 tok/s 95–253 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	10.3 GB	Q8_0	Comfortable
158 tok/s 95–253 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	10.3 GB	Q8_0	Comfortable
158 tok/s 95–253 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	10.3 GB	Q8_0	Comfortable
125 tok/s 75–200 · low confidence	CMP 170HX 8 GB NVIDIA	8 GB	1,490 GB/s	Sep 2021	7.2 GB	Q5_K_M	Tight
120 tok/s 72–193 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	10.3 GB	Q8_0	Comfortable
120 tok/s 72–193 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	10.3 GB	Q8_0	Comfortable
107 tok/s 64–171 · low confidence	CMP 170HX 10 GB NVIDIA	10 GB	1,560 GB/s	Sep 2021	8.2 GB	Q6_K	Tight
100 tok/s 60–161 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	10.3 GB	Q8_0	Comfortable
98.2 tok/s 59–157 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	10.3 GB	Q8_0	Comfortable
96.0 tok/s 58–154 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	10.3 GB	Q8_0	Comfortable
96.0 tok/s 58–154 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	10.3 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: Princeton University,University of Virginia
Organisation type: Academia,Academia
Country: United States of America
Published: 1 November 2024
Authors: Yu Meng, Mengzhou Xia, Danqi Chen

What it does

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

Domain: Language
Task: Language modeling/generation, Question answering
Base model: Gemma 2 9B
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: 9B
Training data: tokens
Epochs: 1

Training compute

The arithmetic performed to train the model, measured in floating-point operations. It is a measure of what the training run cost, not of how fast the finished model answers you.

How it was established: Hardware
Fine-tuning compute: 1.4 × 10¹⁹ FLOP

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 H100 SXM5 80GB
Chips used: 8
Wall-clock time: 2 hours
Power draw: 11.0 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 (restricted use)
Training code: Open source
Hugging Face: princeton-nlp

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: SimPO: Simple Preference Optimization with a Reference-Free Reward
Last updated: 28 November 2025

The extremes

The ten fastest GPUs for SimPO

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 SimPO

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

Quadro 6000

Memory needed

5.1 GB

Fastest

376 tok/s

SimPO is small enough at 9B parameters that hardware is rarely the obstacle — 582 of the cards we track can run it, including cards several years old.

The entry point is the Quadro 6000: 6 GB of memory, Q3_K_M compression, roughly 15.5 tokens per second.

At the other end, a B200 generates roughly 376 tokens per second on it, on the strength of 8,000 GB/s of memory bandwidth.

What this model is

SimPO was published by Princeton University,University of Virginia, in United States of America, in November 2024. academia,Academia is the category the publisher falls under.

It works in Language, and is recorded as doing language modeling/generation, Question answering.

Its starting point was Gemma 2 9B — most models at this scale are adapted from an existing base rather than built from nothing.

The weights are published, so it can be downloaded and run on your own hardware indefinitely, offline, with no account attached. It is published under the princeton-nlp organisation on Hugging Face.

What decides the speed

The median result is around 21.1 tokens per second; 541 cards produce text faster than most people read it.

Being dense, it reads all of itself per token, which is why the ordering by speed below follows the ordering by memory bandwidth so closely.

Without the attention layout on record, the memory column is an approximation. It is close enough to choose hardware by, and least reliable at long context.

Step by step

How to choose a GPU for SimPO

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

01

Check what it needs before anything else

Every card here has been checked against SimPO — around 5.1 GB at Q3_K_M. Capacity is the gate — a card either holds it or it does not.
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 SimPO.
03

Choose how far you will compress it

The quantisation column varies by card, because a bigger card holds a more accurate copy of SimPO — Q3_K_M on the smallest card that fits. Set a floor to hold the comparison at one level.
04

Compare tokens per second, not specifications

Sort by speed to see how cards rank for SimPO. It will not match a gaming ordering — generation is bound by memory bandwidth, which is why the B200 tops it at 376 tok/s.
05

Check the fit verdict before buying

Tight means SimPO loads and works, with no room to raise the context later. Comfortable means you can. The difference matters more than a few tokens per second.
06

Check the card from the other side

Each card page repeats this sweep for every model we hold. It answers what else the hardware is good for, beyond SimPO.

Answers

SimPO — common questions

Is SimPO open source?

Its weights are published, so SimPO 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.

How many parameters does SimPO have?

SimPO has 9B parameters. 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.

Who created SimPO?

SimPO was published by Princeton University,University of Virginia, based in United States of America, categorised as academia,Academia.

When was SimPO released?

SimPO was published in November 2024. Capability per parameter has improved considerably since, so a newer model of the same size is often the better use of the same hardware.

What is SimPO used for?

SimPO works in Language, and is recorded as handling language modeling/generation, Question answering. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

Where can I download SimPO?

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

Can I run SimPO if it does not fit in my GPU?

It can be split between the card and system memory, but SimPO generates painfully slowly that way — the nearest miss we calculate is short by 1.6 GB. Nothing on this page assumes offloading.

Would two GPUs run SimPO faster?

A second card roughly doubles the memory available but not the generation rate. With 582 cards already able to run SimPO alone, the case for pairing is weak.

Why does the quantisation differ between cards for SimPO?

A larger card holds a more accurate copy. Across the cards that run SimPO, 4 compression levels are used; the floor control above pins it to one.

How accurate are these SimPO speed estimates?

These are estimates with real error bars. The fastest result here, 226–602 tok/s on the B200, could reasonably land anywhere in its published range depending on which runtime you use.

What GPU do I need to run SimPO?

The smallest card in our catalogue that holds SimPO is the Quadro 6000, with 6 GB of memory. It runs the model at Q3_K_M using about 5.1 GB, and produces roughly 15.5 tokens per second. 582 cards in total can run it.

How fast is SimPO on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 376 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 541 of the cards that can run SimPO clear that.

How much VRAM does SimPO need?

About 5.1 GB at Q3_K_M compression, 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.

Can I run SimPO on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q5_K_M, using about 7.2 GB and generating roughly 125 tokens per second — a tight fit.

Can I run SimPO on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 10.3 GB and generating roughly 42.9 tokens per second — a tight fit.

Can I run SimPO on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 10.3 GB and generating roughly 53.2 tokens per second — a comfortable fit.

Can I run SimPO on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 10.3 GB and generating roughly 63.1 tokens per second — a comfortable fit.

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