SGPT BE 5.8B TPS calculator

Open weights Peking University 5.8B parameters August 2022

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 · Q3_K_M · 17.2 tok/s

Fastest card

B200

584 tok/s · 180 GB

Which GPUs can run SGPT BE 5.8B?

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
584 tok/s 351–935 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	6.9 GB	Q8_0	Comfortable
584 tok/s 351–935 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	6.9 GB	Q8_0	Comfortable
466 tok/s 280–746 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	6.9 GB	Q8_0	Comfortable
466 tok/s 280–746 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	6.9 GB	Q8_0	Comfortable
373 tok/s 224–597 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	6.9 GB	Q8_0	Comfortable
357 tok/s 214–571 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	6.9 GB	Q8_0	Comfortable
357 tok/s 214–571 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	6.9 GB	Q8_0	Comfortable
342 tok/s 205–547 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	6.9 GB	Q8_0	Comfortable
303 tok/s 182–485 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	6.9 GB	Q8_0	Comfortable
303 tok/s 182–485 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	6.9 GB	Q8_0	Comfortable
303 tok/s 182–485 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	6.9 GB	Q8_0	Comfortable
288 tok/s 173–460 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	6.9 GB	Q8_0	Comfortable
245 tok/s 147–393 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	6.9 GB	Q8_0	Comfortable
245 tok/s 147–393 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	6.9 GB	Q8_0	Comfortable
245 tok/s 147–393 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	6.9 GB	Q8_0	Comfortable
245 tok/s 147–393 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	6.9 GB	Q8_0	Comfortable
245 tok/s 147–393 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	6.9 GB	Q8_0	Comfortable
187 tok/s 112–299 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	6.9 GB	Q8_0	Comfortable
187 tok/s 112–299 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	6.9 GB	Q8_0	Comfortable
156 tok/s 93–249 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	6.9 GB	Q8_0	Comfortable
152 tok/s 91–244 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	6.9 GB	Q8_0	Comfortable
149 tok/s 89–238 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	6.9 GB	Q8_0	Comfortable
149 tok/s 89–238 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	6.9 GB	Q8_0	Comfortable
149 tok/s 89–238 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	6.9 GB	Q8_0	Comfortable
149 tok/s 89–238 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	6.9 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: Peking University
Organisation type: Academia
Country: China
Published: 5 August 2022
Authors: Niklas Muennighoff

What it does

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

Domain: Language
Task: Semantic search, Semantic embedding, Entity embedding
Base model: GPT-J-6B

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: 5.8B
Training data: tokens

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: 1.6 × 10²⁰ 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 A100 SXM4 40 GB
Chips used: 8
Wall-clock time: 60 hours
Power draw: 6.4 kW
Cloud vendor: Oracle

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: Muennighoff

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: SGPT: GPT Sentence Embeddings for Semantic Search
Last updated: 28 November 2025

The extremes

The ten fastest GPUs that run SGPT BE 5.8B

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 SGPT BE 5.8B

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

What you need to run it

Minimum card

Tesla C1080

Memory needed

3.5 GB

Fastest

584 tok/s

SGPT BE 5.8B is small enough at 5.8B parameters that hardware is rarely the obstacle — 818 of the cards we track can run it, including cards several years old.

The least hardware that works is a Tesla C1080. Its 4 GB is enough at Q3_K_M compression, giving roughly 17.2 tokens per second.

At the other end, a B200 generates roughly 584 tokens per second on it, on the strength of 8,000 GB/s of memory bandwidth.

Background

SGPT BE 5.8B was published by Peking University, in China, in August 2022. It comes out of academia.

It works in Language, and is recorded as doing semantic search, Semantic embedding, Entity embedding.

Its starting point was GPT-J-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. It is published under the Muennighoff organisation on Hugging Face.

Reading the throughput figures

The median result is around 24.7 tokens per second; 746 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.

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.

Step by step

How to choose a GPU for SGPT BE 5.8B

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

01

Start from the memory column

Look at what SGPT BE 5.8B actually needs — around 3.5 GB at Q3_K_M. No amount of processing power compensates for a card that cannot hold it.
02

Match the context to your actual use

Set the context to what you will actually use. The cache grows with the conversation, and it is the usual reason SGPT BE 5.8B stops fitting a card that seemed fine.
03

Decide how much compression you will accept

Each card runs the least-compressed copy it can hold — Q3_K_M on the smallest card that fits. Setting a floor drops the cards that only manage SGPT BE 5.8B by squeezing it further than you would want.
04

Rank by throughput rather than spec sheet

Sort by speed to see how cards rank for SGPT BE 5.8B. It will not match a gaming ordering — generation is bound by memory bandwidth, which is why the B200 tops it at 584 tok/s.
05

Look at the headroom, not just the fit

The fit column separates cards that just manage SGPT BE 5.8B from those with room to spare. Buy for the second if the context might grow.
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 SGPT BE 5.8B is settled.

Answers

SGPT BE 5.8B — common questions

What GPU do I need to run SGPT BE 5.8B?

The smallest card in our catalogue that holds SGPT BE 5.8B is the Tesla C1080, with 4 GB of memory. It runs the model at Q3_K_M using about 3.5 GB, and produces roughly 17.2 tokens per second. 818 cards in total can run it.

How fast is SGPT BE 5.8B on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 584 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 746 of the cards that can run SGPT BE 5.8B clear that.

How much VRAM does SGPT BE 5.8B need?

About 3.5 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 SGPT BE 5.8B on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q8_0, using about 6.9 GB and generating roughly 109 tokens per second — a tight fit.

Can I run SGPT BE 5.8B on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 6.9 GB and generating roughly 66.6 tokens per second — a comfortable fit.

Can I run SGPT BE 5.8B on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 6.9 GB and generating roughly 82.5 tokens per second — a comfortable fit.

Can I run SGPT BE 5.8B on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 6.9 GB and generating roughly 97.9 tokens per second — a comfortable fit.

Is SGPT BE 5.8B open source?

Its weights are published, so SGPT BE 5.8B 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 SGPT BE 5.8B have?

SGPT BE 5.8B has 5.8B parameters. 5.8B. 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 SGPT BE 5.8B?

SGPT BE 5.8B was published by Peking University, based in China, categorised as academia.

When was SGPT BE 5.8B released?

SGPT BE 5.8B was published in August 2022. 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 SGPT BE 5.8B used for?

SGPT BE 5.8B works in Language, and is recorded as handling semantic search, Semantic embedding, Entity embedding. These are the areas it was designed around; they describe intent rather than a hard boundary.

Where can I download SGPT BE 5.8B?

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

Can I run SGPT BE 5.8B 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 SGPT BE 5.8B assume it is fully resident.

Would two GPUs run SGPT BE 5.8B faster?

Capacity adds across cards; throughput does not. Since 818 of the cards we track already hold SGPT BE 5.8B on their own, a second card is rarely the answer here.

Why does the quantisation differ between cards for SGPT BE 5.8B?

Each card is shown running the least-compressed copy it can hold, and SGPT BE 5.8B appears at 4 different compression levels across the cards that fit it. Bigger cards get the more accurate version.

How accurate are these SGPT BE 5.8B speed estimates?

Every figure is derived from memory bandwidth and model size, not benchmarked. That is why each is published as a range such as 351–935 tok/s on the B200 rather than a single number.

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