AlphaGeometry TPS calculator

Open weights Google DeepMind,New York University (NYU) 151M parameters January 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

818 cards that can run it

818 cards we hold specifications for

Smallest card that fits

Tesla C1080

4 GB · Q8_0 · 244 tok/s

Fastest card

B200

22,439 tok/s · 180 GB

Which GPUs can run AlphaGeometry?

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
22,439 tok/s 13,463–35,902 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	0.9 GB	Q8_0	Comfortable
22,439 tok/s 13,463–35,902 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	0.9 GB	Q8_0	Comfortable
17,918 tok/s 10,751–28,669 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	0.9 GB	Q8_0	Comfortable
17,918 tok/s 10,751–28,669 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	0.9 GB	Q8_0	Comfortable
14,330 tok/s 8,598–22,928 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	0.9 GB	Q8_0	Comfortable
13,716 tok/s 8,229–21,945 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.9 GB	Q8_0	Comfortable
13,716 tok/s 8,229–21,945 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.9 GB	Q8_0	Comfortable
13,127 tok/s 7,876–21,003 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	0.9 GB	Q8_0	Comfortable
11,650 tok/s 6,990–18,640 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	0.9 GB	Q8_0	Comfortable
11,650 tok/s 6,990–18,640 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	0.9 GB	Q8_0	Comfortable
11,650 tok/s 6,990–18,640 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	0.9 GB	Q8_0	Comfortable
11,051 tok/s 6,631–17,682 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	0.9 GB	Q8_0	Comfortable
9,424 tok/s 5,655–15,079 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.9 GB	Q8_0	Comfortable
9,424 tok/s 5,655–15,079 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	0.9 GB	Q8_0	Comfortable
9,424 tok/s 5,655–15,079 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	0.9 GB	Q8_0	Comfortable
9,424 tok/s 5,655–15,079 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.9 GB	Q8_0	Comfortable
9,424 tok/s 5,655–15,079 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	0.9 GB	Q8_0	Comfortable
7,176 tok/s 4,306–11,481 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	0.9 GB	Q8_0	Comfortable
7,176 tok/s 4,306–11,481 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	0.9 GB	Q8_0	Comfortable
5,980 tok/s 3,588–9,568 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	0.9 GB	Q8_0	Comfortable
5,852 tok/s 3,511–9,364 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	0.9 GB	Q8_0	Comfortable
5,722 tok/s 3,433–9,155 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	0.9 GB	Q8_0	Comfortable
5,722 tok/s 3,433–9,155 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	0.9 GB	Q8_0	Comfortable
5,722 tok/s 3,433–9,155 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	0.9 GB	Q8_0	Comfortable
5,722 tok/s 3,433–9,155 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	0.9 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: Google DeepMind,New York University (NYU)
Organisation type: Industry,Academia
Country: United States of America
Published: 17 January 2024
Authors: Trieu H. Trinh, Yuhuai Wu, Quoc V. Le, He He, Thang Luong

What it does

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

Domain: Mathematics
Task: Geometry, Mathematical reasoning

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: 151M
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: Google TPU v3

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 (unrestricted)
Training code: Open source

How it is classified

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

Why it is tracked: SOTA improvement
Record confidence: Confident
Citations: 574

Sources

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

Reference: Solving olympiad geometry without human demonstrations
Last updated: 1 January 2026

The extremes

The ten fastest GPUs that run AlphaGeometry

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 AlphaGeometry

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

0.9 GB

Fastest

22,439 tok/s

AlphaGeometry is small enough at 151M parameters that hardware is rarely the obstacle — 818 of the cards we track can run it, including cards several years old.

The entry point is the Tesla C1080: 4 GB of memory, Q8_0 compression, roughly 244 tokens per second.

Top of the range is the B200, at roughly 22,439 tokens per second thanks to 8,000 GB/s of bandwidth.

Where it came from

AlphaGeometry was published by Google DeepMind,New York University (NYU), in United States of America, in January 2024. industry,Academia is the category the publisher falls under.

It works in Mathematics, and is recorded as doing geometry, Mathematical reasoning.

The weights being open is what puts this page in the calculator rather than only in the catalogue: it is a model you can actually hold.

Understanding the speeds

The median result is around 630.1 tokens per second; 818 cards produce text faster than most people read it.

It is a dense model, so every parameter is read for every token produced. That makes speed track memory bandwidth almost exactly — a card with twice the bandwidth generates roughly twice as fast.

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.

What went into building it

Its inclusion criterion is sOTA improvement.

Step by step

How to choose a GPU for AlphaGeometry

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 AlphaGeometry — around 0.9 GB at Q8_0. 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 AlphaGeometry.
03

Choose how far you will compress it

Compression is what makes AlphaGeometry fit smaller cards, at some cost in accuracy — Q8_0 on the smallest card that fits. A minimum quality removes the ones that go too far.
04

Rank by throughput rather than spec sheet

The speed ordering for AlphaGeometry is effectively an ordering by memory bandwidth, which is why the B200 tops it at 22,439 tok/s.
05

Look at the headroom, not just the fit

A tight fit runs AlphaGeometry but leaves nothing spare for a longer conversation; comfortable has headroom. If you expect to grow the context, buy for comfortable.
06

Open the card you have settled on

Following a card through to its own page shows every other model it can hold, which is the question that follows once AlphaGeometry is settled.

Answers

AlphaGeometry — common questions

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

It can be split between the card and system memory, but AlphaGeometry generates painfully slowly that way. Nothing on this page assumes offloading.

Would two GPUs run AlphaGeometry faster?

Two cards buy memory rather than speed. That matters for AlphaGeometry only if one card cannot hold it — 818 can, so a second adds little.

Why does the quantisation differ between cards for AlphaGeometry?

Each card is shown running the least-compressed copy it can hold, and AlphaGeometry appears at 1 different compression levels across the cards that fit it. Bigger cards get the more accurate version.

How accurate are these AlphaGeometry speed estimates?

These are estimates with real error bars. The fastest result here, 13,463–35,902 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 AlphaGeometry?

The smallest card in our catalogue that holds AlphaGeometry is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 0.9 GB, and produces roughly 244 tokens per second. 818 cards in total can run it.

How fast is AlphaGeometry on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 22,439 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 818 of the cards that can run AlphaGeometry clear that.

How much VRAM does AlphaGeometry need?

About 0.9 GB at Q8_0 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 AlphaGeometry on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q8_0, using about 0.9 GB and generating roughly 4,179 tokens per second — a comfortable fit.

Can I run AlphaGeometry on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 0.9 GB and generating roughly 2,559 tokens per second — a comfortable fit.

Can I run AlphaGeometry on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 0.9 GB and generating roughly 3,169 tokens per second — a comfortable fit.

Can I run AlphaGeometry on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 0.9 GB and generating roughly 3,758 tokens per second — a comfortable fit.

Is AlphaGeometry open source?

Its weights are published, so AlphaGeometry 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 AlphaGeometry have?

AlphaGeometry has 151M parameters. "Overall, the transformer has 151 million parameters, excluding embedding layers at its input and output heads.". 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 AlphaGeometry?

AlphaGeometry was published by Google DeepMind,New York University (NYU), based in United States of America, categorised as industry,Academia.

When was AlphaGeometry released?

AlphaGeometry was published in January 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 AlphaGeometry used for?

AlphaGeometry works in Mathematics, and is recorded as handling geometry, Mathematical reasoning. These are the areas it was designed around; they describe intent rather than a hard boundary.

Where can I download AlphaGeometry?

The weights for AlphaGeometry are published, though we do not hold a repository link for it. This site calculates hardware requirements rather than hosting model files.

Source

Original publication

Record last updated 1 January 2026

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