BLUUMI TPS calculator

Open weights University of Turku,Hugging Face 176B parameters November 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

17 cards that can run it

818 cards we hold specifications for

Smallest card that fits

Radeon Instinct MI250

128 GB · Q4_K_M · 14.2 tok/s

Fastest card

Radeon Instinct MI300

28.4 tok/s · 128 GB

Which GPUs can run BLUUMI?

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

17 cards match

Calculating

					Needs	Quantisation	Fit
28.4 tok/s 17–45 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	107.2 GB	Q4_K_M	Tight
28.0 tok/s 17–45 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	148.2 GB	Q6_K	Tight
27.2 tok/s 16–43 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	107.2 GB	Q4_K_M	Tight
27.2 tok/s 16–43 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	107.2 GB	Q4_K_M	Tight
23.1 tok/s 14–37 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	107.2 GB	Q4_K_M	Tight
19.3 tok/s 12–31 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	189.1 GB	Q8_0	Comfortable
15.4 tok/s 9–25 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	189.1 GB	Q8_0	Comfortable
15.4 tok/s 9–25 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	189.1 GB	Q8_0	Comfortable
14.5 tok/s 9–23 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	148.2 GB	Q6_K	Tight
14.5 tok/s 9–23 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	148.2 GB	Q6_K	Tight
14.2 tok/s 9–23 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	107.2 GB	Q4_K_M	Tight
14.2 tok/s 9–23 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	107.2 GB	Q4_K_M	Tight
11.8 tok/s 7–19 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	107.2 GB	Q4_K_M	Tight
11.6 tok/s 7–19 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	107.2 GB	Q4_K_M	Tight
11.3 tok/s 7–18 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	189.1 GB	Q8_0	Comfortable
1.5 tok/s 1–2 · low confidence	GB10 NVIDIA	128 GB	273 GB/s	Oct 2025	107.2 GB	Q4_K_M	Tight
1.5 tok/s 1–2 · low confidence	Jetson T5000 NVIDIA	128 GB	273 GB/s	Aug 2025	107.2 GB	Q4_K_M	Tight

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: University of Turku,Hugging Face
Organisation type: Academia,Industry
Country: Finland, United States of America
Published: 3 November 2023
Authors: Risto Luukkonen, Ville Komulainen, Jouni Luoma, Anni Eskelinen, Jenna Kanerva, Hanna-Mari Kupari, Filip Ginter, Veronika Laippala, Niklas Muennighoff, Aleksandra Piktus, Thomas Wang, Nouamane Tazi, Teven Le Scao, Thomas Wolf, Osma Suominen, Samuli Sairanen, Mikko Merioksa, Jyrki Heinonen, Aija Vahtola, Samuel Antao, Sampo Pyysalo

What it does

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

Domain: Language
Task: Language modeling
Base model: BLOOM-176B

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: 176B
Training data: 38,000,000,000 tokens
Epochs: 8
Batch size: 4,194,304

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: AMD Radeon Instinct MI250X
Data centre: LUMI Supercomputer

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: Unreleased
Hugging Face: TurkuNLP

How it is classified

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

Foundation model: Yes
Why it is tracked: SOTA improvement
Record confidence: Likely
Citations: 70

Sources

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

Reference: FinGPT: Large Generative Models for a Small Language
Last updated: 25 May 2026

The extremes

The ten fastest GPUs that run BLUUMI

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 BLUUMI

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

Radeon Instinct MI250

Memory needed

107.2 GB

Fastest

28.4 tok/s

BLUUMI reaches a parameter count of 176B. That puts it above consumer hardware, into the range where a card is bought for this purpose rather than repurposed for it. The number of cards we track that can hold it: 17.

At the low end it is handled by Radeon Instinct MI250, with a memory capacity of 128 GB, running it at a compression of Q4_K_M and producing around 14.2 tokens per second.

The quickest result comes from Radeon Instinct MI300, generating roughly 28.4 tokens per second on the strength of a memory bandwidth of 6,550 GB/s.

About this model

BLUUMI was published by University of Turku,Hugging Face, in the country recorded as Finland, during November 2023. The category the publisher falls under is academia,Industry.

It works in the domain of Language, and is recorded as performing the task of language modeling.

Rather than being trained from scratch, it is derived from BLOOM-176B. That is the usual way a specialised model is produced.

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. On Hugging Face it is published under the organisation TurkuNLP.

How fast it runs, and why

Across every card that can run it, the middle of the range sits at 14.5 tokens per second. Producing text faster than most people read it: 15 of them.

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

It was trained on a corpus of about 38,000,000,000 tokens of text.

The reason it appears in this catalogue at all: sOTA improvement.

Step by step

How to choose a GPU for BLUUMI

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

The table lists every card able to hold BLUUMI, needing around 107.2 GB at a compression of Q4_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 BLUUMI.
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 Q4_K_M 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 BLUUMI. It will not match a gaming ordering, because generation is bound by memory bandwidth. The card topping the list is Radeon Instinct MI300, at 28.4 tok/s.
05

Look at the headroom, not just the fit

The fit column separates cards that just manage it from those with room to spare, in the case of BLUUMI. 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

Check the card from the other side

Every card name links to its own page, which runs the same calculation across the whole model catalogue. A card is usually bought for more than one model, so it is worth a look before buying for BLUUMI.

Answers

BLUUMI — common questions

BLUUMI— 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. The nearest miss we calculate falls short by 20.8 GB. Every figure here assumes the whole model is resident on the card.

BLUUMI— would two GPUs run it faster?

Capacity adds across cards; throughput does not. The number of cards already holding it on their own: 17. So a second card is rarely the answer here.

BLUUMI— why does the quantisation differ between cards?

A larger card holds a more accurate copy. The number of compression levels used across the cards that run it: 3. Bigger cards get the more accurate version, and the quality floor above pins the comparison to one level.

BLUUMI— how accurate are these speed estimates?

Every figure is derived from memory bandwidth and model size, not benchmarked, which is why each is published as a range rather than a single number. One example: 17–45 tok/s on Radeon Instinct MI300. The same model and card vary by thirty to fifty per cent depending on the inference software and its version.

BLUUMI— what GPU do I need to run it?

The smallest card in our catalogue that holds it is Radeon Instinct MI250, with a memory capacity of 128 GB. It runs the model at a compression of Q4_K_M using about 107.2 GB, and produces roughly 14.2 tokens per second. The number of cards able to run it in total: 17.

BLUUMI— how fast is it on a GPU?

It depends on the card. The quickest we calculate is Radeon Instinct MI300, at about 28.4 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: 15.

BLUUMI— how much VRAM does it need?

It needs about 107.2 GB at a compression of Q4_K_M, 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.

BLUUMI— 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.

BLUUMI— how many parameters does it have?

It has a parameter count of 176B. 176 billion. 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.

BLUUMI— who created it?

It was published by University of Turku,Hugging Face, based in Finland, an organisation categorised as academia,Industry.

BLUUMI— when was it released?

It was published in November 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.

BLUUMI— what is it used for?

It works in the domain of Language, and is recorded as handling the task of language modeling. 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.

BLUUMI— where can I download it?

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

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