DeepSeek LLM 67B TPS calculator

Open weights DeepSeek 67B 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

61 cards that can run it

818 cards we hold specifications for

Smallest card that fits

A100 PCIe 40 GB

40 GB · IQ4_XS · 24.2 tok/s

Fastest card

B200

50.6 tok/s · 180 GB

Which GPUs can run DeepSeek LLM 67B?

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

61 cards match

Calculating

					Needs	Quantisation	Fit
50.6 tok/s 43–61	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	69.7 GB	Q8_0	Comfortable
50.6 tok/s 43–61	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	69.7 GB	Q8_0	Comfortable
40.4 tok/s 24–65 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	69.7 GB	Q8_0	Comfortable
40.4 tok/s 24–65 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	69.7 GB	Q8_0	Comfortable
32.3 tok/s 19–52 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	69.7 GB	Q8_0	Comfortable
30.9 tok/s 26–37	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	69.7 GB	Q8_0	Comfortable
30.9 tok/s 26–37	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	69.7 GB	Q8_0	Comfortable
29.6 tok/s 18–47 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	69.7 GB	Q8_0	Comfortable
27.3 tok/s 23–33	GRID A100B NVIDIA	48 GB	1,870 GB/s	May 2020	38.5 GB	Q4_K_M	Tight
26.3 tok/s 16–42 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	69.7 GB	Q8_0	Comfortable
26.3 tok/s 16–42 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	69.7 GB	Q8_0	Comfortable
26.3 tok/s 16–42 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	69.7 GB	Q8_0	Comfortable
24.9 tok/s 21–30	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	69.7 GB	Q8_0	Comfortable
24.2 tok/s 21–29	A100 PCIe 40 GB NVIDIA	40 GB	1,560 GB/s	Jun 2020	34.6 GB	IQ4_XS	Tight
24.2 tok/s 21–29	A100 SXM4 40 GB NVIDIA	40 GB	1,560 GB/s	May 2020	34.6 GB	IQ4_XS	Tight
24.2 tok/s 21–29	A800 PCIe 40 GB NVIDIA	40 GB	1,560 GB/s	Nov 2022	34.6 GB	IQ4_XS	Tight
21.2 tok/s 18–25	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	69.7 GB	Q8_0	Comfortable
21.2 tok/s 18–25	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	69.7 GB	Q8_0	Tight
21.2 tok/s 18–25	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	69.7 GB	Q8_0	Comfortable
21.2 tok/s 18–25	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	69.7 GB	Q8_0	Comfortable
21.2 tok/s 18–25	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	69.7 GB	Q8_0	Tight
19.6 tok/s 17–23	RTX PRO 5000 Blackwell NVIDIA	48 GB	1,340 GB/s	Mar 2025	38.5 GB	Q4_K_M	Tight
18.6 tok/s 16–22	H100 SXM5 64 GB NVIDIA	64 GB	2,020 GB/s	Mar 2023	54.1 GB	Q6_K	Tight
16.2 tok/s 10–26 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	69.7 GB	Q8_0	Comfortable
16.2 tok/s 10–26 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	69.7 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: DeepSeek
Organisation type: Industry
Country: China
Published: 5 January 2024
Authors: Xiao Bi, Deli Chen, Guanting Chen, Shanhuang Chen, Damai Dai, Chengqi Deng, Honghui Ding, Kai Dong, Qiushi Du, Zhe Fu, Huazuo Gao, Kaige Gao, Wenjun Gao, Ruiqi Ge, Kang Guan, Daya Guo, Jianzhong Guo, Guangbo Hao, Zhewen Hao, Ying He, Wenjie Hu, Panpan Huang, Erhang Li, Guowei Li, Jiashi Li, Yao Li, Y.K. Li, Wenfeng Liang, Fangyun Lin, A.X. Liu, Bo Liu, Wen Liu, Xiaodong Liu, Xin Liu, Yiyuan Liu, H…

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

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: 67B
Training data: 2,000,000,000,000 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: 8 × 10²³ FLOP
How it was established: Operation counting

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: deepseek-ai

How it is classified

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

Likely above 10²³ FLOP: Yes
Record confidence: Confident

Sources

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

Reference: DeepSeek LLM: Scaling Open-Source Language Models with Longtermism
Last updated: 28 November 2025

The extremes

The ten fastest GPUs that run DeepSeek LLM 67B

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 DeepSeek LLM 67B

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

A100 PCIe 40 GB

Memory needed

34.6 GB

Fastest

50.6 tok/s

DeepSeek LLM 67B reaches a parameter count of 67B. 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: 61.

At the low end it is handled by A100 PCIe 40 GB, with a memory capacity of 40 GB, running it at a compression of IQ4_XS and producing around 24.2 tokens per second.

The fastest we calculate for it is B200, generating roughly 50.6 tokens per second on the strength of a memory bandwidth of 8,000 GB/s.

Background

DeepSeek LLM 67B was published by DeepSeek, in the country recorded as China, during January 2024. The category the publisher falls under is industry.

It works in the domain of Language, and is recorded as performing the task of chat, Language modeling/generation, Question answering.

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 deepseek-ai.

Reading the throughput figures

Half the cards that hold it manage more than 12.9 tokens per second. Producing text faster than most people read it: 51 of them.

Every weight participates in every token here, so bandwidth is the whole story: the ranking below is effectively a ranking of memory throughput.

Because the architecture is recorded, the memory column is derived rather than estimated.

Training and provenance

Training it took a computation budget of roughly 8 × 10²³ FLOP. That figure measures what producing the model cost, and has no bearing on how fast it answers.

It was trained on a corpus of about 2,000,000,000,000 tokens of text.

Step by step

How to choose a GPU for DeepSeek LLM 67B

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 DeepSeek LLM 67B, needing around 34.6 GB at a compression of IQ4_XS. No amount of processing power compensates for a card that cannot hold it.
02

Match the context to your actual use

Longer conversations cost memory on top of what the weights need. Move the slider to your real working length before trusting any row for DeepSeek LLM 67B.
03

Decide how much compression you will accept

Each card runs the least-compressed copy it can hold, reaching a compression of IQ4_XS 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

The speed ordering is effectively an ordering by memory bandwidth, for DeepSeek LLM 67B. It will not match a gaming ordering, because generation is bound by memory bandwidth. The card topping the list is B200, at 50.6 tok/s.
05

Look at the headroom, not just the fit

A tight fit runs, but leaves nothing spare for a longer conversation, in the case of DeepSeek LLM 67B. 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

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 you have settled on DeepSeek LLM 67B.

Answers

DeepSeek LLM 67B — common questions

DeepSeek LLM 67B— how fast is it on a GPU?

It depends on the card. The quickest we calculate is B200, at about 50.6 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: 51.

DeepSeek LLM 67B— how much VRAM does it need?

It needs about 34.6 GB at a compression of IQ4_XS, 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.

DeepSeek LLM 67B— 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.

DeepSeek LLM 67B— how many parameters does it have?

It has a parameter count of 67B. 67B. 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.

DeepSeek LLM 67B— who created it?

It was published by DeepSeek, based in China, an organisation categorised as industry.

DeepSeek LLM 67B— when was it released?

It 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.

DeepSeek LLM 67B— what is it used for?

It works in the domain of Language, and is recorded as handling the task of chat, Language modeling/generation, Question answering. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

DeepSeek LLM 67B— where can I download it?

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

DeepSeek LLM 67B— how much compute was used to train it?

Training consumed around 8 × 10²³ 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.

DeepSeek LLM 67B— 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 9.7 GB. Every figure here assumes the whole model is resident on the card.

DeepSeek LLM 67B— would two GPUs run it faster?

A second card roughly doubles the memory available but not the generation rate. The number already able to run it alone: 61. So a second card is rarely the answer here.

DeepSeek LLM 67B— why does the quantisation differ between cards?

Because capacity varies, so does how hard it has to be squeezed. The number of distinct levels in the table above: 4. Bigger cards get the more accurate version, and the quality floor above pins the comparison to one level.

DeepSeek LLM 67B— how accurate are these speed estimates?

They are calculated from specifications rather than measured, and each carries a range. One example: 43–61 tok/s on B200. The same model and card vary by thirty to fifty per cent depending on the inference software and its version.

DeepSeek LLM 67B— what GPU do I need to run it?

The smallest card in our catalogue that holds it is A100 PCIe 40 GB, with a memory capacity of 40 GB. It runs the model at a compression of IQ4_XS using about 34.6 GB, and produces roughly 24.2 tokens per second. The number of cards able to run it in total: 61.

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