Ling-lite-1.5 ("Bailing") TPS calculator

Open weights Ant Group 16.8B parameters March 2025

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 of 818 cards that can run it

Smallest card that fits

P102-101

10 GB · Q3_K_M · 18.5 tok/s

Fastest card

B200

202 tok/s · 180 GB

Which GPUs can run Ling-lite-1.5 ("Bailing")?

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
202 tok/s 121–323 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	18.7 GB	Q8_0	Comfortable
202 tok/s 121–323 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	18.7 GB	Q8_0	Comfortable
161 tok/s 97–258 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	18.7 GB	Q8_0	Comfortable
161 tok/s 97–258 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	18.7 GB	Q8_0	Comfortable
129 tok/s 77–206 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	18.7 GB	Q8_0	Comfortable
123 tok/s 74–197 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	18.7 GB	Q8_0	Comfortable
123 tok/s 74–197 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	18.7 GB	Q8_0	Comfortable
118 tok/s 71–189 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	18.7 GB	Q8_0	Comfortable
106 tok/s 64–170 · low confidence	CMP 170HX 10 GB NVIDIA	10 GB	1,560 GB/s	Sep 2021	8.9 GB	Q3_K_M	Tight
105 tok/s 63–168 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	18.7 GB	Q8_0	Comfortable
105 tok/s 63–168 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	18.7 GB	Q8_0	Comfortable
105 tok/s 63–168 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	18.7 GB	Q8_0	Comfortable
99.3 tok/s 60–159 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	18.7 GB	Q8_0	Comfortable
84.7 tok/s 51–136 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	18.7 GB	Q8_0	Comfortable
84.7 tok/s 51–136 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	18.7 GB	Q8_0	Comfortable
84.7 tok/s 51–136 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	18.7 GB	Q8_0	Comfortable
84.7 tok/s 51–136 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	18.7 GB	Q8_0	Comfortable
84.7 tok/s 51–136 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	18.7 GB	Q8_0	Comfortable
64.5 tok/s 39–103 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	18.7 GB	Q8_0	Comfortable
64.5 tok/s 39–103 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	18.7 GB	Q8_0	Comfortable
56.5 tok/s 34–90 · low confidence	GeForce RTX 3080 12 GB NVIDIA	12 GB	912 GB/s	Jan 2022	9.9 GB	IQ4_XS	Tight
56.5 tok/s 34–90 · low confidence	GeForce RTX 3080 Ti NVIDIA	12 GB	912 GB/s	May 2021	9.9 GB	IQ4_XS	Tight
53.8 tok/s 32–86 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	18.7 GB	Q8_0	Comfortable
52.6 tok/s 32–84 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	18.7 GB	Q8_0	Comfortable
51.7 tok/s 31–83 · low confidence	CMP 90HX NVIDIA	10 GB	760 GB/s	Jul 2021	8.9 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: Ant Group
Organisation type: Industry
Country: China
Published: 10 March 2025
Authors: Ling Team, Binwei Zeng, Chao Huang, Chao Zhang, Changxin Tian, Cong Chen, Dingnan Jin, Feng Yu, Feng Zhu, Feng Yuan, Fakang Wang, Gangshan Wang, Guangyao Zhai, Haitao Zhang, Huizhong Li, Jun Zhou, Jia Liu, Junpeng Fang, Junjie Ou, Jun Hu, Ji Luo, Ji Zhang, Jian Liu, Jian Sha, Jianxue Qian, Jiewei Wu, Junping Zhao, Jianguo Li, Jubao Feng, Jingchao Di, Junming Xu, Jinghua Yao, Kuan Xu, Kewei Du, Lon…

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, Code generation

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: 16.8B
Training data: 9,000,000,000,000 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: 1.5 × 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 (unrestricted)
Training code: Unreleased
Hugging Face: inclusionAI

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: Every FLOP Counts: Scaling a 300B Mixture-of-Experts LING LLM without Premium GPUs
Last updated: 11 February 2026

The extremes

The ten fastest GPUs for Ling-lite-1.5 ("Bailing")

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 Ling-lite-1.5 ("Bailing")

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

Hardware requirements in practice

Minimum card

P102-101

Memory needed

8.9 GB

Fastest

202 tok/s

Ling-lite-1.5 ("Bailing") is small enough at 16.8B parameters that hardware is rarely the obstacle — 306 of the cards we track can run it, including cards several years old.

The entry point is the P102-101: 10 GB of memory, Q3_K_M compression, roughly 18.5 tokens per second.

The quickest result comes from a B200 at around 202 tokens per second — its 8,000 GB/s of bandwidth is what buys that.

What this model is

Ling-lite-1.5 ("Bailing") was published by Ant Group, in China, in March 2025. industry is the category the publisher falls under.

It works in Language, and is recorded as doing language modeling/generation, Question answering, 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 inclusionAI organisation on Hugging Face.

What decides the speed

The median result is around 20.8 tokens per second; 268 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.

Without the attention layout on record, the memory column is an approximation. It is close enough to choose hardware by, and least reliable at long context.

What went into building it

Producing it required around 1.5 × 10²³ FLOP of arithmetic, which is a statement about the training budget rather than about inference.

It was trained on about 9,000,000,000,000 tokens of text.

Step by step

How to choose a GPU for Ling-lite-1.5 ("Bailing")

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

The table lists every card that can hold Ling-lite-1.5 ("Bailing") — around 8.9 GB at Q3_K_M. That figure, not the card's headline performance, is what decides whether it runs.
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 Ling-lite-1.5 ("Bailing") stops fitting a card that seemed fine.
03

Decide how much compression you will accept

The quantisation column varies by card, because a bigger card holds a more accurate copy of Ling-lite-1.5 ("Bailing") — Q3_K_M on the smallest card that fits. Set a floor to hold the comparison at one level.
04

Sort by speed

Ranking by tokens per second for Ling-lite-1.5 ("Bailing") follows memory bandwidth, not core counts, which is why the B200 tops it at 202 tok/s.
05

Read the fit column last

Tight means Ling-lite-1.5 ("Bailing") 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 Ling-lite-1.5 ("Bailing").

Answers

Ling-lite-1.5 ("Bailing") — common questions

Is Ling-lite-1.5 ("Bailing") open source?

Its weights are published, so Ling-lite-1.5 ("Bailing") 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 Ling-lite-1.5 ("Bailing") have?

Ling-lite-1.5 ("Bailing") has 16.8B parameters. 16.8 billion parameters with 2.75 billion activated parameters. 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 Ling-lite-1.5 ("Bailing")?

Ling-lite-1.5 ("Bailing") was published by Ant Group, based in China, categorised as industry.

When was Ling-lite-1.5 ("Bailing") released?

Ling-lite-1.5 ("Bailing") was published in March 2025.

What is Ling-lite-1.5 ("Bailing") used for?

Ling-lite-1.5 ("Bailing") works in Language, and is recorded as handling language modeling/generation, Question answering, Code generation. These are the areas it was designed around; they describe intent rather than a hard boundary.

Where can I download Ling-lite-1.5 ("Bailing")?

Its weights are published under the inclusionAI 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 Ling-lite-1.5 ("Bailing")?

Around 1.5 × 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.

Can I run Ling-lite-1.5 ("Bailing") 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 Ling-lite-1.5 ("Bailing") is rarely worth using — the nearest miss we calculate is short by 3.7 GB. Every figure here assumes the whole model is on the card.

Would two GPUs run Ling-lite-1.5 ("Bailing") faster?

Two cards buy memory rather than speed. That matters for Ling-lite-1.5 ("Bailing") only if one card cannot hold it — 306 can, so a second adds little.

Why does the quantisation differ between cards for Ling-lite-1.5 ("Bailing")?

Because capacity varies, so does how hard Ling-lite-1.5 ("Bailing") has to be squeezed — 5 distinct levels appear in the table above. Set a minimum quality to compare at one.

How accurate are these Ling-lite-1.5 ("Bailing") 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 121–323 tok/s on the B200 rather than a single number.

What GPU do I need to run Ling-lite-1.5 ("Bailing")?

The smallest card in our catalogue that holds Ling-lite-1.5 ("Bailing") is the P102-101, with 10 GB of memory. It runs the model at Q3_K_M using about 8.9 GB, and produces roughly 18.5 tokens per second. 306 cards in total can run it.

How fast is Ling-lite-1.5 ("Bailing") on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 202 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 268 of the cards that can run Ling-lite-1.5 ("Bailing") clear that.

How much VRAM does Ling-lite-1.5 ("Bailing") need?

About 8.9 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 Ling-lite-1.5 ("Bailing") on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at IQ4_XS, using about 9.9 GB and generating roughly 56.5 tokens per second — a tight fit.

Can I run Ling-lite-1.5 ("Bailing") on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q5_K_M, using about 12.8 GB and generating roughly 50.9 tokens per second — a tight fit.

Can I run Ling-lite-1.5 ("Bailing") on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 18.7 GB and generating roughly 33.8 tokens per second — a tight fit.

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