OpenThaiGPT v1.0.0 (7B) TPS calculator

Open weights Mahidol University,AI Entrepreneurs Association of Thailand 6.8B parameters April 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

589 cards that can run it

818 cards we hold specifications for

Smallest card that fits

Tesla K20c

5 GB · IQ4_XS · 27.0 tok/s

Fastest card

B200

498 tok/s · 180 GB

Which GPUs can run OpenThaiGPT v1.0.0 (7B)?

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

589 cards match

Calculating

					Needs	Quantisation	Fit
498 tok/s 299–796 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	8.0 GB	Q8_0	Comfortable
498 tok/s 299–796 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	8.0 GB	Q8_0	Comfortable
397 tok/s 238–636 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	8.0 GB	Q8_0	Comfortable
397 tok/s 238–636 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	8.0 GB	Q8_0	Comfortable
318 tok/s 191–508 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	8.0 GB	Q8_0	Comfortable
304 tok/s 182–487 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	8.0 GB	Q8_0	Comfortable
304 tok/s 182–487 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	8.0 GB	Q8_0	Comfortable
291 tok/s 175–466 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	8.0 GB	Q8_0	Comfortable
258 tok/s 155–413 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	8.0 GB	Q8_0	Comfortable
258 tok/s 155–413 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	8.0 GB	Q8_0	Comfortable
258 tok/s 155–413 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	8.0 GB	Q8_0	Comfortable
245 tok/s 147–392 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	8.0 GB	Q8_0	Comfortable
209 tok/s 125–334 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	8.0 GB	Q8_0	Comfortable
209 tok/s 125–334 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	8.0 GB	Q8_0	Comfortable
209 tok/s 125–334 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	8.0 GB	Q8_0	Comfortable
209 tok/s 125–334 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	8.0 GB	Q8_0	Comfortable
209 tok/s 125–334 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	8.0 GB	Q8_0	Comfortable
159 tok/s 95–255 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	8.0 GB	Q8_0	Comfortable
159 tok/s 95–255 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	8.0 GB	Q8_0	Comfortable
135 tok/s 81–215 · low confidence	CMP 170HX 8 GB NVIDIA	8 GB	1,490 GB/s	Sep 2021	6.4 GB	Q6_K	Tight
133 tok/s 80–212 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	8.0 GB	Q8_0	Comfortable
130 tok/s 78–208 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	8.0 GB	Q8_0	Comfortable
127 tok/s 76–203 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	8.0 GB	Q8_0	Comfortable
127 tok/s 76–203 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	8.0 GB	Q8_0	Comfortable
127 tok/s 76–203 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	8.0 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: Mahidol University,AI Entrepreneurs Association of Thailand
Organisation type: Academia
Country: Thailand
Published: 8 April 2024
Authors: Kobkrit Viriyayudhakorn, Sumeth Yuenyong, Thaweewat Rugsujarit, Jillaphat Jaroenkantasima, Norapat Buppodom, Koravich Sangkaew, Peerawat Rojratchadakorn, Surapon Nonesung, Chanon Utupon, Sadhis Wongprayoon, Nucharee Thongthungwong, Chawakorn Phiantham, Patteera Triamamornwooth. Nattarika Juntarapaoraya, Kriangkrai Saetan, Pitikorn Khlaisamniang

What it does

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

Domain: Language
Task: Chat, Language modeling/generation, Question answering, Retrieval-augmented generation
Base model: Llama 2-7B

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: 6.8B
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.

Training compute: 2.7 × 10²¹ FLOP
How it was established: Operation counting
Fine-tuning compute: 4.1 × 10¹⁶ FLOP

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: openthaigpt

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: OpenThaiGPT 7b 1.0.0
Last updated: 11 February 2026

The extremes

The ten fastest GPUs that run OpenThaiGPT v1.0.0 (7B)

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 OpenThaiGPT v1.0.0 (7B)

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 it takes to run this model

Minimum card

Tesla K20c

Memory needed

4.4 GB

Fastest

498 tok/s

OpenThaiGPT v1.0.0 (7B) is small enough at 6.8B parameters that hardware is rarely the obstacle — 589 of the cards we track can run it, including cards several years old.

The smallest card that holds it is the Tesla K20c with 5 GB, running it at IQ4_XS and producing around 27.0 tokens per second.

A B200 is the fastest we calculate for it: about 498 tokens per second, from 8,000 GB/s of memory bandwidth.

What this model is

OpenThaiGPT v1.0.0 (7B) was published by Mahidol University,AI Entrepreneurs Association of Thailand, in Thailand, in April 2024. academia is the category the publisher falls under.

It works in Language, and is recorded as doing chat, Language modeling/generation, Question answering, Retrieval-augmented generation.

It builds on Llama 2-7B, which is why it shares that model's general shape and size.

Published weights mean the model runs on your machine rather than someone else's, which is what makes the hardware question below answerable at all. It is published under the openthaigpt organisation on Hugging Face.

What decides the speed

Half the cards that hold it manage more than 26.9 tokens per second, and 562 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.

Its internal architecture is not on file, so memory is approximated from the parameter count and marked accordingly. Expect the real figure to differ, more so at long context.

How it was trained

Producing it required around 2.7 × 10²¹ FLOP of arithmetic, which is a statement about the training budget rather than about inference.

Step by step

How to choose a GPU for OpenThaiGPT v1.0.0 (7B)

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 OpenThaiGPT v1.0.0 (7B) — around 4.4 GB at IQ4_XS. Capacity is the gate — a card either holds it or it does not.
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 OpenThaiGPT v1.0.0 (7B) 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 — IQ4_XS on the smallest card that fits. Setting a floor drops the cards that only manage OpenThaiGPT v1.0.0 (7B) by squeezing it further than you would want.
04

Sort by speed

The speed ordering for OpenThaiGPT v1.0.0 (7B) is effectively an ordering by memory bandwidth, which is why the B200 tops it at 498 tok/s.
05

Check the fit verdict before buying

A tight fit runs OpenThaiGPT v1.0.0 (7B) 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

Every card name links to its own page, which runs the same calculation across the whole model catalogue. Worth a look before buying for OpenThaiGPT v1.0.0 (7B) alone — a card is usually bought for more than one model.

Answers

OpenThaiGPT v1.0.0 (7B) — common questions

What is OpenThaiGPT v1.0.0 (7B) used for?

OpenThaiGPT v1.0.0 (7B) works in Language, and is recorded as handling chat, Language modeling/generation, Question answering, Retrieval-augmented generation. 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.

Where can I download OpenThaiGPT v1.0.0 (7B)?

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

How much compute was used to train OpenThaiGPT v1.0.0 (7B)?

Around 2.7 × 10²¹ FLOP. 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 OpenThaiGPT v1.0.0 (7B) if it does not fit in my GPU?

It can be split between the card and system memory, but OpenThaiGPT v1.0.0 (7B) generates painfully slowly that way — the nearest miss we calculate is short by 1.2 GB. Nothing on this page assumes offloading.

Would two GPUs run OpenThaiGPT v1.0.0 (7B) faster?

Two cards buy memory rather than speed. That matters for OpenThaiGPT v1.0.0 (7B) only if one card cannot hold it — 589 can, so a second adds little.

Why does the quantisation differ between cards for OpenThaiGPT v1.0.0 (7B)?

Because capacity varies, so does how hard OpenThaiGPT v1.0.0 (7B) has to be squeezed — 4 distinct levels appear in the table above. Set a minimum quality to compare at one.

How accurate are these OpenThaiGPT v1.0.0 (7B) speed estimates?

These are estimates with real error bars. The fastest result here, 299–796 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 OpenThaiGPT v1.0.0 (7B)?

The smallest card in our catalogue that holds OpenThaiGPT v1.0.0 (7B) is the Tesla K20c, with 5 GB of memory. It runs the model at IQ4_XS using about 4.4 GB, and produces roughly 27.0 tokens per second. 589 cards in total can run it.

How fast is OpenThaiGPT v1.0.0 (7B) on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 498 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 562 of the cards that can run OpenThaiGPT v1.0.0 (7B) clear that.

How much VRAM does OpenThaiGPT v1.0.0 (7B) need?

About 4.4 GB at IQ4_XS 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 OpenThaiGPT v1.0.0 (7B) on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q6_K, using about 6.4 GB and generating roughly 135 tokens per second — a tight fit.

Can I run OpenThaiGPT v1.0.0 (7B) on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 8.0 GB and generating roughly 56.7 tokens per second — a comfortable fit.

Can I run OpenThaiGPT v1.0.0 (7B) on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 8.0 GB and generating roughly 70.3 tokens per second — a comfortable fit.

Can I run OpenThaiGPT v1.0.0 (7B) on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 8.0 GB and generating roughly 83.3 tokens per second — a comfortable fit.

Is OpenThaiGPT v1.0.0 (7B) open source?

Its weights are published, so OpenThaiGPT v1.0.0 (7B) 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 OpenThaiGPT v1.0.0 (7B) have?

OpenThaiGPT v1.0.0 (7B) has 6.8B parameters. https://huggingface.co/openthaigpt/openthaigpt-1.0.0-7b-chat. 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 OpenThaiGPT v1.0.0 (7B)?

OpenThaiGPT v1.0.0 (7B) was published by Mahidol University,AI Entrepreneurs Association of Thailand, based in Thailand, categorised as academia.

When was OpenThaiGPT v1.0.0 (7B) released?

OpenThaiGPT v1.0.0 (7B) was published in April 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.

Source

Original publication

Record last updated 11 February 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