Qwen2.5-14B TPS calculator

Open weights Alibaba 14.7B parameters September 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

306 cards that can run it

818 cards we hold specifications for

Smallest card that fits

P102-101

10 GB · Q3_K_M · 21.2 tok/s

Fastest card

B200

230 tok/s · 180 GB

Which GPUs can run Qwen2.5-14B?

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

306 cards match

Calculating

					Needs	Quantisation	Fit
230 tok/s 196–277	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	17.0 GB	Q8_0	Comfortable
230 tok/s 196–277	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	17.0 GB	Q8_0	Comfortable
184 tok/s 110–294 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	17.0 GB	Q8_0	Comfortable
184 tok/s 110–294 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	17.0 GB	Q8_0	Comfortable
147 tok/s 88–236 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	17.0 GB	Q8_0	Comfortable
141 tok/s 120–169	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	17.0 GB	Q8_0	Comfortable
141 tok/s 120–169	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	17.0 GB	Q8_0	Comfortable
135 tok/s 81–216 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	17.0 GB	Q8_0	Comfortable
121 tok/s 103–146	CMP 170HX 10 GB NVIDIA	10 GB	1,560 GB/s	Sep 2021	8.5 GB	Q3_K_M	Tight
120 tok/s 72–191 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	17.0 GB	Q8_0	Comfortable
120 tok/s 72–191 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	17.0 GB	Q8_0	Comfortable
120 tok/s 72–191 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	17.0 GB	Q8_0	Comfortable
114 tok/s 96–136	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	17.0 GB	Q8_0	Comfortable
96.8 tok/s 82–116	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	17.0 GB	Q8_0	Comfortable
96.8 tok/s 82–116	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	17.0 GB	Q8_0	Comfortable
96.8 tok/s 82–116	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	17.0 GB	Q8_0	Comfortable
96.8 tok/s 82–116	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	17.0 GB	Q8_0	Comfortable
96.8 tok/s 82–116	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	17.0 GB	Q8_0	Comfortable
73.7 tok/s 44–118 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	17.0 GB	Q8_0	Comfortable
73.7 tok/s 44–118 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	17.0 GB	Q8_0	Comfortable
61.4 tok/s 37–98 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	17.0 GB	Q8_0	Comfortable
60.7 tok/s 52–73	GeForce RTX 3080 12 GB NVIDIA	12 GB	912 GB/s	Jan 2022	10.2 GB	Q4_K_M	Tight
60.7 tok/s 52–73	GeForce RTX 3080 Ti NVIDIA	12 GB	912 GB/s	May 2021	10.2 GB	Q4_K_M	Tight
60.1 tok/s 36–96 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	17.0 GB	Q8_0	Comfortable
59.1 tok/s 50–71	CMP 90HX NVIDIA	10 GB	760 GB/s	Jul 2021	8.5 GB	Q3_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: Alibaba
Organisation type: Industry
Country: China
Published: 19 September 2024
Authors: Qwen Team

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, Quantitative reasoning

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: 14.7B
Training data: 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: 1.6 × 10²⁴ 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 (unrestricted)
Training code: Unreleased
Hugging Face: Qwen

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: Qwen2.5: A Party of Foundation Models!
Last updated: 28 November 2025

The extremes

The ten fastest GPUs that run Qwen2.5-14B

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 Qwen2.5-14B

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

P102-101

Memory needed

8.5 GB

Fastest

230 tok/s

Qwen2.5-14B is small enough at 14.7B parameters that hardware is rarely the obstacle — 306 of the cards we track can run it, including cards several years old.

At the low end, a P102-101 handles it — 10 GB, at Q3_K_M, for about 21.2 tokens per second.

Top of the range is the B200, at roughly 230 tokens per second thanks to 8,000 GB/s of bandwidth.

Background

Qwen2.5-14B was published by Alibaba, in China, in September 2024. The organisation is categorised as industry.

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

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 Qwen organisation on Hugging Face.

Reading the throughput figures

The median result is around 21.5 tokens per second; 266 cards produce text faster than most people read it.

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.

What went into building it

The training run consumed about 1.6 × 10²⁴ FLOP. That figure describes the cost of creating it and has no bearing on how quickly it generates text.

Step by step

How to choose a GPU for Qwen2.5-14B

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

Look at what Qwen2.5-14B actually needs — around 8.5 GB at Q3_K_M. No amount of processing power compensates for a card that cannot hold it.
02

Set the context length you will work at

The conversation occupies memory too, and grows as it goes. Set the slider to the length you expect: at long context Qwen2.5-14B can slip off a card that handles short questions easily.
03

Set a quality floor

The quantisation column varies by card, because a bigger card holds a more accurate copy of Qwen2.5-14B — Q3_K_M on the smallest card that fits. Set a floor to hold the comparison at one level.
04

Sort by speed

The speed ordering for Qwen2.5-14B is effectively an ordering by memory bandwidth, which is why the B200 tops it at 230 tok/s.
05

Read the fit column last

Tight means Qwen2.5-14B 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 Qwen2.5-14B.

Answers

Qwen2.5-14B — common questions

Who created Qwen2.5-14B?

Qwen2.5-14B was published by Alibaba, based in China, categorised as industry.

When was Qwen2.5-14B released?

Qwen2.5-14B was published in September 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.

What is Qwen2.5-14B used for?

Qwen2.5-14B works in Language, and is recorded as handling language modeling/generation, Question answering, Quantitative reasoning. These are the areas it was designed around; they describe intent rather than a hard boundary.

Where can I download Qwen2.5-14B?

Its weights are published under the Qwen 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 Qwen2.5-14B?

Around 1.6 × 10²⁴ 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 Qwen2.5-14B 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 — the nearest miss we calculate is short by 3.0 GB. Our figures for Qwen2.5-14B assume it is fully resident.

Would two GPUs run Qwen2.5-14B faster?

Two cards buy memory rather than speed. That matters for Qwen2.5-14B only if one card cannot hold it — 306 can, so a second adds little.

Why does the quantisation differ between cards for Qwen2.5-14B?

Because capacity varies, so does how hard Qwen2.5-14B has to be squeezed — 5 distinct levels appear in the table above. Set a minimum quality to compare at one.

How accurate are these Qwen2.5-14B speed estimates?

These are estimates with real error bars. The fastest result here, 196–277 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 Qwen2.5-14B?

The smallest card in our catalogue that holds Qwen2.5-14B is the P102-101, with 10 GB of memory. It runs the model at Q3_K_M using about 8.5 GB, and produces roughly 21.2 tokens per second. 306 cards in total can run it.

How fast is Qwen2.5-14B on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 230 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 266 of the cards that can run Qwen2.5-14B clear that.

How much VRAM does Qwen2.5-14B need?

About 8.5 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 Qwen2.5-14B on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q4_K_M, using about 10.2 GB and generating roughly 60.7 tokens per second — a tight fit.

Can I run Qwen2.5-14B on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q6_K, using about 13.6 GB and generating roughly 47.3 tokens per second — a tight fit.

Can I run Qwen2.5-14B on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 17.0 GB and generating roughly 38.6 tokens per second — a comfortable fit.

Is Qwen2.5-14B open source?

Its weights are published, so Qwen2.5-14B 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 Qwen2.5-14B have?

Qwen2.5-14B has 14.7B parameters. 14.7B according to the model card https://qwenlm.github.io/blog/qwen2.5-llm/. 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.

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