Pythia-160m TPS calculator

Open weights EleutherAI,Booz Allen Hamilton, McLean,University of Cambridge,Indraprastha Institute of Information Technology Delhi,Stability AI,datasaur.ai,University of Amsterdam 160M parameters April 2023

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 · 230 tok/s

Fastest card

B200

21,176 tok/s · 180 GB

Which GPUs can run Pythia-160m?

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
21,176 tok/s 12,706–33,882 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	0.9 GB	Q8_0	Comfortable
21,176 tok/s 12,706–33,882 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	0.9 GB	Q8_0	Comfortable
16,910 tok/s 10,146–27,056 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	0.9 GB	Q8_0	Comfortable
16,910 tok/s 10,146–27,056 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	0.9 GB	Q8_0	Comfortable
13,524 tok/s 8,114–21,638 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	0.9 GB	Q8_0	Comfortable
12,944 tok/s 7,766–20,711 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.9 GB	Q8_0	Comfortable
12,944 tok/s 7,766–20,711 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	0.9 GB	Q8_0	Comfortable
12,388 tok/s 7,433–19,821 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	0.9 GB	Q8_0	Comfortable
10,995 tok/s 6,597–17,591 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	0.9 GB	Q8_0	Comfortable
10,995 tok/s 6,597–17,591 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	0.9 GB	Q8_0	Comfortable
10,995 tok/s 6,597–17,591 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	0.9 GB	Q8_0	Comfortable
10,429 tok/s 6,258–16,687 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	0.9 GB	Q8_0	Comfortable
8,894 tok/s 5,336–14,231 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.9 GB	Q8_0	Comfortable
8,894 tok/s 5,336–14,231 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	0.9 GB	Q8_0	Comfortable
8,894 tok/s 5,336–14,231 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	0.9 GB	Q8_0	Comfortable
8,894 tok/s 5,336–14,231 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	0.9 GB	Q8_0	Comfortable
8,894 tok/s 5,336–14,231 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	0.9 GB	Q8_0	Comfortable
6,772 tok/s 4,063–10,836 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	0.9 GB	Q8_0	Comfortable
6,772 tok/s 4,063–10,836 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	0.9 GB	Q8_0	Comfortable
5,644 tok/s 3,386–9,030 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	0.9 GB	Q8_0	Comfortable
5,523 tok/s 3,314–8,837 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	0.9 GB	Q8_0	Comfortable
5,400 tok/s 3,240–8,640 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	0.9 GB	Q8_0	Comfortable
5,400 tok/s 3,240–8,640 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	0.9 GB	Q8_0	Comfortable
5,400 tok/s 3,240–8,640 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	0.9 GB	Q8_0	Comfortable
5,400 tok/s 3,240–8,640 · 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: EleutherAI,Booz Allen Hamilton, McLean,University of Cambridge,Indraprastha Institute of Information Technology Delhi,Stability AI,datasaur.ai,University of Amsterdam
Organisation type: Research collective,Industry,Academia,Academia,Industry,Industry,Academia
Country: United States of America, United Kingdom of Great Britain and Northern Ireland, India, Netherlands
Published: 3 April 2023
Authors: Stella Biderman, Hailey Schoelkopf, Quentin Anthony, Herbie Bradley, Kyle O'Brien, Eric Hallahan, Mohammad Aflah Khan, Shivanshu Purohit, USVSN Sai Prashanth, Edward Raff, Aviya Skowron, Lintang Sutawika, Oskar van der Wal

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

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: 160M
Training data: 299,892,736,000 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.

Training compute: 2.9 × 10²⁰ FLOP
How it was established: Operation counting,Hardware

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 SXM4 40 GB
Chips used: 32
Chip-hours: 1,030
Power draw: 25.5 kW
Cloud vendor: Amazon Web Services
Data centre: AWS US East

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.

Record confidence: Confident
Citations: 1,862
Benchmark data: Pythia-160m

Sources

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

Reference: Pythia: A Suite for Analyzing Large Language Models Across Training and Scaling
Last updated: 25 May 2026

The extremes

The ten fastest GPUs that run Pythia-160m

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 Pythia-160m

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

The hardware side

Minimum card

Tesla C1080

Memory needed

0.9 GB

Fastest

21,176 tok/s

Pythia-160m is small enough at 160M parameters that hardware is rarely the obstacle — 818 of the cards we track can run it, including cards several years old.

At the low end, a Tesla C1080 handles it — 4 GB, at Q8_0, for about 230 tokens per second.

At the other end, a B200 generates roughly 21,176 tokens per second on it, on the strength of 8,000 GB/s of memory bandwidth.

Where it came from

Pythia-160m was published by EleutherAI,Booz Allen Hamilton, McLean,University of Cambridge,Indraprastha Institute of Information Technology Delhi,Stability AI,datasaur.ai,University of Amsterdam, in United States of America, in April 2023. The organisation is categorised as research collective,Industry,Academia,Academia,Industry,Industry,Academia.

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

The weights are published, so it can be downloaded and run on your own hardware indefinitely, offline, with no account attached.

Understanding the speeds

Half the cards that hold it manage more than 594.6 tokens per second, and 818 exceed reading speed outright.

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.

Memory here is estimated from size rather than computed from the architecture, which is not recorded for this model — the numbers are indicative rather than exact.

What went into building it

The training run consumed about 2.9 × 10²⁰ FLOP, on NVIDIA A100 SXM4 40 GB. That figure describes the cost of creating it and has no bearing on how quickly it generates text.

The training set ran to roughly 299,892,736,000 tokens.

Step by step

How to choose a GPU for Pythia-160m

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

Look at what Pythia-160m actually needs — around 0.9 GB at Q8_0. No amount of processing power compensates for a card that cannot hold it.
02

Decide how long your conversations run

Set the context to what you will actually use. The cache grows with the conversation, and it is the usual reason Pythia-160m stops fitting a card that seemed fine.
03

Choose how far you will compress it

Each card runs the least-compressed copy it can hold — Q8_0 on the smallest card that fits. Setting a floor drops the cards that only manage Pythia-160m by squeezing it further than you would want.
04

Sort by speed

Sort by speed to see how cards rank for Pythia-160m. It will not match a gaming ordering — generation is bound by memory bandwidth, which is why the B200 tops it at 21,176 tok/s.
05

Read the fit column last

Tight means Pythia-160m 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 Pythia-160m.

Answers

Pythia-160m — common questions

How accurate are these Pythia-160m speed estimates?

They are calculated from specifications rather than measured, and each carries a range — 12,706–33,882 tok/s on the B200, for instance. The same model and card vary by thirty to fifty per cent depending on the inference software and its version.

What GPU do I need to run Pythia-160m?

The smallest card in our catalogue that holds Pythia-160m is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 0.9 GB, and produces roughly 230 tokens per second. 818 cards in total can run it.

How fast is Pythia-160m on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 21,176 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 Pythia-160m clear that.

How much VRAM does Pythia-160m 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 Pythia-160m 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 3,944 tokens per second — a comfortable fit.

Can I run Pythia-160m 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 2,415 tokens per second — a comfortable fit.

Can I run Pythia-160m 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 2,991 tokens per second — a comfortable fit.

Can I run Pythia-160m 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 3,547 tokens per second — a comfortable fit.

Is Pythia-160m open source?

Its weights are published, so Pythia-160m 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 Pythia-160m have?

Pythia-160m has 160M parameters. See Table 1 for non-embedding 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 Pythia-160m?

Pythia-160m was published by EleutherAI,Booz Allen Hamilton, McLean,University of Cambridge,Indraprastha Institute of Information Technology Delhi,Stability AI,datasaur.ai,University of Amsterdam, based in United States of America, categorised as research collective,Industry,Academia,Academia,Industry,Industry,Academia.

When was Pythia-160m released?

Pythia-160m was published in April 2023. 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 Pythia-160m used for?

Pythia-160m 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 Pythia-160m?

The weights for Pythia-160m are published, though we do not hold a repository link for it. This site calculates hardware requirements rather than hosting model files.

How much compute was used to train Pythia-160m?

Around 2.9 × 10²⁰ FLOP, on NVIDIA A100 SXM4 40 GB. That measures what producing the model cost and says nothing about how quickly it answers once trained — inference speed comes from memory bandwidth, not from the training budget.

Can I run Pythia-160m if it does not fit in my GPU?

Only by offloading, which is usually a false economy: the part in system memory drags the whole thing down. Our figures for Pythia-160m assume it is fully resident.

Would two GPUs run Pythia-160m faster?

A second card roughly doubles the memory available but not the generation rate. With 818 cards already able to run Pythia-160m alone, the case for pairing is weak.

Why does the quantisation differ between cards for Pythia-160m?

A larger card holds a more accurate copy. Across the cards that run Pythia-160m, 1 compression levels are used; the floor control above pins it to one.

Source

Original publication

Record last updated 25 May 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