CodeT5-large TPS calculator

Open weights Salesforce 770M parameters July 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 · Q8_0 · 47.9 tok/s

Fastest card

B200

4,400 tok/s · 180 GB

Which GPUs can run CodeT5-large?

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
4,400 tok/s 2,640–7,040 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	1.5 GB	Q8_0	Comfortable
4,400 tok/s 2,640–7,040 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	1.5 GB	Q8_0	Comfortable
3,514 tok/s 2,108–5,622 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	1.5 GB	Q8_0	Comfortable
3,514 tok/s 2,108–5,622 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	1.5 GB	Q8_0	Comfortable
2,810 tok/s 1,686–4,496 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	1.5 GB	Q8_0	Comfortable
2,690 tok/s 1,614–4,304 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.5 GB	Q8_0	Comfortable
2,690 tok/s 1,614–4,304 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	1.5 GB	Q8_0	Comfortable
2,574 tok/s 1,545–4,119 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	1.5 GB	Q8_0	Comfortable
2,285 tok/s 1,371–3,655 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	1.5 GB	Q8_0	Comfortable
2,285 tok/s 1,371–3,655 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	1.5 GB	Q8_0	Comfortable
2,285 tok/s 1,371–3,655 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	1.5 GB	Q8_0	Comfortable
2,167 tok/s 1,300–3,467 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	1.5 GB	Q8_0	Comfortable
1,848 tok/s 1,109–2,957 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.5 GB	Q8_0	Comfortable
1,848 tok/s 1,109–2,957 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	1.5 GB	Q8_0	Comfortable
1,848 tok/s 1,109–2,957 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	1.5 GB	Q8_0	Comfortable
1,848 tok/s 1,109–2,957 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	1.5 GB	Q8_0	Comfortable
1,848 tok/s 1,109–2,957 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	1.5 GB	Q8_0	Comfortable
1,407 tok/s 844–2,252 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	1.5 GB	Q8_0	Comfortable
1,407 tok/s 844–2,252 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	1.5 GB	Q8_0	Comfortable
1,173 tok/s 704–1,876 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	1.5 GB	Q8_0	Comfortable
1,148 tok/s 689–1,836 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	1.5 GB	Q8_0	Comfortable
1,122 tok/s 673–1,795 · low confidence	A100 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Nov 2020	1.5 GB	Q8_0	Comfortable
1,122 tok/s 673–1,795 · low confidence	A100X NVIDIA	80 GB	2,040 GB/s	Jun 2021	1.5 GB	Q8_0	Comfortable
1,122 tok/s 673–1,795 · low confidence	A800 SXM4 80 GB NVIDIA	80 GB	2,040 GB/s	Aug 2022	1.5 GB	Q8_0	Comfortable
1,122 tok/s 673–1,795 · low confidence	H100 CNX NVIDIA	80 GB	2,040 GB/s	Mar 2023	1.5 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: Salesforce
Organisation type: Industry
Country: United States of America
Published: 5 July 2022
Authors: Hung Le, Yue Wang, Akhilesh Deepak Gotmare, Silvio Savarese, Steven C.H. Hoi

What it does

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

Domain: Language
Task: Code generation
Numerical format: FP16

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: 770M
Training data: 10,500,000,000 tokens
Epochs: 150

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: Hardware

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
Wall-clock time: 504 hours (21 days)
Compute cost: $4,478

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: Open source
Hugging Face: Salesforce

How it is classified

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

Why it is tracked: SOTA improvement
Record confidence: Likely
Citations: 465

Sources

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

Reference: CodeRL: Mastering Code Generation through Pretrained Models and Deep Reinforcement Learning
Last updated: 25 May 2026

The extremes

The ten fastest GPUs that run CodeT5-large

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 CodeT5-large

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

Tesla C1080

Memory needed

1.5 GB

Fastest

4,400 tok/s

CodeT5-large is small enough at 770M parameters that hardware is rarely the obstacle — 818 of the cards we track can run it, including cards several years old.

At the low end, a Tesla C1080 handles it — 4 GB, at Q8_0, for about 47.9 tokens per second.

At the other end, a B200 generates roughly 4,400 tokens per second on it, on the strength of 8,000 GB/s of memory bandwidth.

What this model is

CodeT5-large was published by Salesforce, in United States of America, in July 2022. The organisation is categorised as industry.

It works in Language, and is recorded as doing code generation.

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 Salesforce organisation on Hugging Face.

What decides the speed

The median result is around 123.6 tokens per second; 809 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.

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.

What went into building it

Producing it required around 2.7 × 10²¹ FLOP of arithmetic, on NVIDIA A100, which is a statement about the training budget rather than about inference.

Around 10,500,000,000 tokens went into training it.

It is tracked in the underlying dataset for one reason in particular: sOTA improvement.

Step by step

How to choose a GPU for CodeT5-large

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

The table lists every card that can hold CodeT5-large — around 1.5 GB at Q8_0. That figure, not the card's headline performance, is what decides whether it runs.
02

Match the context to your actual use

The conversation occupies memory too, and grows as it goes. Set the slider to the length you expect: at long context CodeT5-large can slip off a card that handles short questions easily.
03

Choose how far you will compress it

Each card runs the least-compressed copy it can hold — Q8_0 on the smallest card that fits. Setting a floor drops the cards that only manage CodeT5-large by squeezing it further than you would want.
04

Compare tokens per second, not specifications

The speed ordering for CodeT5-large is effectively an ordering by memory bandwidth, which is why the B200 tops it at 4,400 tok/s.
05

Look at the headroom, not just the fit

Tight means CodeT5-large 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

Open the card you have settled on

Following a card through to its own page shows every other model it can hold, which is the question that follows once CodeT5-large is settled.

Answers

CodeT5-large — common questions

Can I run CodeT5-large on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q8_0, using about 1.5 GB and generating roughly 622 tokens per second — a comfortable fit.

Can I run CodeT5-large on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 1.5 GB and generating roughly 737 tokens per second — a comfortable fit.

Is CodeT5-large open source?

Its weights are published, so CodeT5-large 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 CodeT5-large have?

CodeT5-large has 770M parameters. "We pretrain a CodeT5-large model (770M) from scratch following T5-large’s architecture". 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 CodeT5-large?

CodeT5-large was published by Salesforce, based in United States of America, categorised as industry.

When was CodeT5-large released?

CodeT5-large was published in July 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 CodeT5-large used for?

CodeT5-large works in Language, and is recorded as handling code generation. These are the areas it was designed around; they describe intent rather than a hard boundary.

Where can I download CodeT5-large?

Its weights are published under the Salesforce 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 CodeT5-large?

Around 2.7 × 10²¹ FLOP, on NVIDIA A100. 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 CodeT5-large if it does not fit in my GPU?

Partly. Layers that do not fit sit in system memory and run at a fraction of the speed, so a mostly-offloaded CodeT5-large is rarely worth using. Every figure here assumes the whole model is on the card.

Would two GPUs run CodeT5-large faster?

A second card roughly doubles the memory available but not the generation rate. With 818 cards already able to run CodeT5-large alone, the case for pairing is weak.

Why does the quantisation differ between cards for CodeT5-large?

Each card is shown running the least-compressed copy it can hold, and CodeT5-large appears at 1 different compression levels across the cards that fit it. Bigger cards get the more accurate version.

How accurate are these CodeT5-large speed estimates?

These are estimates with real error bars. The fastest result here, 2,640–7,040 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 CodeT5-large?

The smallest card in our catalogue that holds CodeT5-large is the Tesla C1080, with 4 GB of memory. It runs the model at Q8_0 using about 1.5 GB, and produces roughly 47.9 tokens per second. 818 cards in total can run it.

How fast is CodeT5-large on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 4,400 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 809 of the cards that can run CodeT5-large clear that.

How much VRAM does CodeT5-large need?

About 1.5 GB at Q8_0 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 CodeT5-large on a 8 GB GPU?

Yes. A CMP 170HX 8 GB with 8 GB runs it at Q8_0, using about 1.5 GB and generating roughly 820 tokens per second — a comfortable fit.

Can I run CodeT5-large on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q8_0, using about 1.5 GB and generating roughly 502 tokens per second — a comfortable fit.

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