Ring-mini-linear-2.0 TPS calculator

Open weights Ant Group 16.4B parameters October 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 · 19.0 tok/s

Fastest card

B200

207 tok/s · 180 GB

Which GPUs can run Ring-mini-linear-2.0?

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
207 tok/s 124–331 · low confidence	B200 NVIDIA	180 GB	8,000 GB/s	Jan 2024	18.3 GB	Q8_0	Comfortable
207 tok/s 124–331 · low confidence	B300 NVIDIA	288 GB	8,000 GB/s	Sep 2025	18.3 GB	Q8_0	Comfortable
165 tok/s 99–264 · low confidence	Radeon Instinct MI350X AMD	288 GB	8,190 GB/s	Jan 2025	18.3 GB	Q8_0	Comfortable
165 tok/s 99–264 · low confidence	Radeon Instinct MI355X AMD	288 GB	8,190 GB/s	Jan 2025	18.3 GB	Q8_0	Comfortable
132 tok/s 79–211 · low confidence	Radeon Instinct MI300 AMD	128 GB	6,550 GB/s	Jan 2023	18.3 GB	Q8_0	Comfortable
126 tok/s 76–202 · low confidence	H200 NVL NVIDIA	141 GB	4,890 GB/s	Nov 2024	18.3 GB	Q8_0	Comfortable
126 tok/s 76–202 · low confidence	H200 SXM 141 GB NVIDIA	141 GB	4,890 GB/s	Nov 2024	18.3 GB	Q8_0	Comfortable
121 tok/s 73–193 · low confidence	Radeon Instinct MI325X AMD	256 GB	6,000 GB/s	Oct 2024	18.3 GB	Q8_0	Comfortable
109 tok/s 65–174 · low confidence	CMP 170HX 10 GB NVIDIA	10 GB	1,560 GB/s	Sep 2021	8.7 GB	Q3_K_M	Tight
107 tok/s 64–172 · low confidence	Radeon Instinct MI300A AMD	128 GB	5,325 GB/s	Dec 2023	18.3 GB	Q8_0	Comfortable
107 tok/s 64–172 · low confidence	Radeon Instinct MI300X AMD	192 GB	5,325 GB/s	Dec 2023	18.3 GB	Q8_0	Comfortable
107 tok/s 64–172 · low confidence	Radeon Instinct MI308X AMD	192 GB	5,325 GB/s	Dec 2023	18.3 GB	Q8_0	Comfortable
102 tok/s 61–163 · low confidence	H100 NVL 94 GB NVIDIA	94 GB	3,940 GB/s	Mar 2023	18.3 GB	Q8_0	Comfortable
86.8 tok/s 52–139 · low confidence	H100 PCIe 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	18.3 GB	Q8_0	Comfortable
86.8 tok/s 52–139 · low confidence	H100 SXM5 80 GB NVIDIA	80 GB	3,360 GB/s	Oct 2022	18.3 GB	Q8_0	Comfortable
86.8 tok/s 52–139 · low confidence	H100 SXM5 94 GB NVIDIA	94 GB	3,360 GB/s	Mar 2023	18.3 GB	Q8_0	Comfortable
86.8 tok/s 52–139 · low confidence	H100 SXM5 96 GB NVIDIA	96 GB	3,360 GB/s	Mar 2023	18.3 GB	Q8_0	Comfortable
86.8 tok/s 52–139 · low confidence	H800 SXM5 NVIDIA	80 GB	3,360 GB/s	Mar 2023	18.3 GB	Q8_0	Comfortable
66.1 tok/s 40–106 · low confidence	Radeon Instinct MI250 AMD	128 GB	3,280 GB/s	Nov 2021	18.3 GB	Q8_0	Comfortable
66.1 tok/s 40–106 · low confidence	Radeon Instinct MI250X AMD	128 GB	3,280 GB/s	Nov 2021	18.3 GB	Q8_0	Comfortable
55.1 tok/s 33–88 · low confidence	Data Center GPU Max 1550 Intel	128 GB	3,280 GB/s	Jan 2023	18.3 GB	Q8_0	Comfortable
54.4 tok/s 33–87 · low confidence	GeForce RTX 3080 12 GB NVIDIA	12 GB	912 GB/s	Jan 2022	10.6 GB	Q4_K_M	Tight
54.4 tok/s 33–87 · low confidence	GeForce RTX 3080 Ti NVIDIA	12 GB	912 GB/s	May 2021	10.6 GB	Q4_K_M	Tight
53.9 tok/s 32–86 · low confidence	Data Center GPU Max Subsystem Intel	128 GB	3,210 GB/s	Jan 2023	18.3 GB	Q8_0	Comfortable
53.0 tok/s 32–85 · low confidence	CMP 90HX NVIDIA	10 GB	760 GB/s	Jul 2021	8.7 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: 23 October 2025
Authors: Bin Han, Caizhi Tang, Chen Liang, Donghao Zhang, Fan Yuan, Feng Zhu, Jie Gao, Jingyu Hu, Longfei Li, Meng Li, Mingyang Zhang, Peijie Jiang, Peng Jiao, Qian Zhao, Qingyuan Yang, Wenbo Shen, Xinxing Yang, Yalin Zhang, Yankun Ren, Yao Zhao, Yibo Cao, Yixuan Sun, Yue Zhang, Yuchen Fang, Zibin Lin, Zixuan Cheng, Jun Zhou

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, Code generation, Mathematical reasoning
Base model: Ling-mini-base-2.0-20T
Numerical format: FP8

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: 16.4B
Training data: 600,000,000,000 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: 1.7 × 10²³ FLOP
How it was established: Operation counting
Fine-tuning compute: 3.4 × 10²¹ FLOP

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 H800 SXM5
Chips used: 288
Power draw: 393.6 kW

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 Attention Matters: An Efficient Hybrid Architecture for Long-Context Reasoning
Last updated: 28 November 2025

The extremes

The ten fastest GPUs for Ring-mini-linear-2.0

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 Ring-mini-linear-2.0

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

8.7 GB

Fastest

207 tok/s

Ring-mini-linear-2.0 is small enough at 16.4B 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 19.0 tokens per second.

Top of the range is the B200, at roughly 207 tokens per second thanks to 8,000 GB/s of bandwidth.

What this model is

Ring-mini-linear-2.0 was published by Ant Group, in China, in October 2025. It comes out of industry.

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

It is derived from Ling-mini-base-2.0-20T rather than trained from scratch, which is the usual way a specialised model is produced.

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 20.7 tokens per second; 269 cards produce text faster than most people read it.

Being dense, it reads all of itself per token, which is why the ordering by speed below follows the ordering by memory bandwidth so closely.

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.

Training and provenance

Producing it required around 1.7 × 10²³ FLOP of arithmetic, on NVIDIA H800 SXM5, which is a statement about the training budget rather than about inference.

It was trained on about 600,000,000,000 tokens of text.

Step by step

How to choose a GPU for Ring-mini-linear-2.0

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 Ring-mini-linear-2.0 actually needs — around 8.7 GB at Q3_K_M. No amount of processing power compensates for a card that cannot hold it.
02

Decide how long your conversations run

Longer conversations cost memory on top of what the weights need. Move the slider to your real working length before trusting any row for Ring-mini-linear-2.0.
03

Decide how much compression you will accept

The quantisation column varies by card, because a bigger card holds a more accurate copy of Ring-mini-linear-2.0 — Q3_K_M on the smallest card that fits. Set a floor to hold the comparison at one level.
04

Rank by throughput rather than spec sheet

The speed ordering for Ring-mini-linear-2.0 is effectively an ordering by memory bandwidth, which is why the B200 tops it at 207 tok/s.
05

Look at the headroom, not just the fit

A tight fit runs Ring-mini-linear-2.0 but leaves nothing spare for a longer conversation; comfortable has headroom. If you expect to grow the context, buy for comfortable.
06

See what else that card runs

Each card page repeats this sweep for every model we hold. It answers what else the hardware is good for, beyond Ring-mini-linear-2.0.

Answers

Ring-mini-linear-2.0 — common questions

Who created Ring-mini-linear-2.0?

Ring-mini-linear-2.0 was published by Ant Group, based in China, categorised as industry.

When was Ring-mini-linear-2.0 released?

Ring-mini-linear-2.0 was published in October 2025.

What is Ring-mini-linear-2.0 used for?

Ring-mini-linear-2.0 works in Language, and is recorded as handling language modeling/generation, Question answering, Quantitative reasoning, Code generation, Mathematical reasoning. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.

Where can I download Ring-mini-linear-2.0?

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 Ring-mini-linear-2.0?

Around 1.7 × 10²³ FLOP, on NVIDIA H800 SXM5. 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 Ring-mini-linear-2.0 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 3.4 GB. Our figures for Ring-mini-linear-2.0 assume it is fully resident.

Would two GPUs run Ring-mini-linear-2.0 faster?

A second card roughly doubles the memory available but not the generation rate. With 306 cards already able to run Ring-mini-linear-2.0 alone, the case for pairing is weak.

Why does the quantisation differ between cards for Ring-mini-linear-2.0?

Because capacity varies, so does how hard Ring-mini-linear-2.0 has to be squeezed — 6 distinct levels appear in the table above. Set a minimum quality to compare at one.

How accurate are these Ring-mini-linear-2.0 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 124–331 tok/s on the B200 rather than a single number.

What GPU do I need to run Ring-mini-linear-2.0?

The smallest card in our catalogue that holds Ring-mini-linear-2.0 is the P102-101, with 10 GB of memory. It runs the model at Q3_K_M using about 8.7 GB, and produces roughly 19.0 tokens per second. 306 cards in total can run it.

How fast is Ring-mini-linear-2.0 on a GPU?

It depends on the card. The quickest we calculate is a B200 at about 207 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 269 of the cards that can run Ring-mini-linear-2.0 clear that.

How much VRAM does Ring-mini-linear-2.0 need?

About 8.7 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 Ring-mini-linear-2.0 on a 12 GB GPU?

Yes. A GeForce RTX 3080 Ti with 12 GB runs it at Q4_K_M, using about 10.6 GB and generating roughly 54.4 tokens per second — a tight fit.

Can I run Ring-mini-linear-2.0 on a 16 GB GPU?

Yes. A Tesla V100 SXM2 16 GB with 16 GB runs it at Q5_K_M, using about 12.5 GB and generating roughly 52.1 tokens per second — a tight fit.

Can I run Ring-mini-linear-2.0 on a 24 GB GPU?

Yes. A GeForce RTX 5090 D V2 with 24 GB runs it at Q8_0, using about 18.3 GB and generating roughly 34.6 tokens per second — a tight fit.

Is Ring-mini-linear-2.0 open source?

Its weights are published, so Ring-mini-linear-2.0 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 Ring-mini-linear-2.0 have?

Ring-mini-linear-2.0 has 16.4B parameters. Ring-mini-linear-2.0 comprises 16B parameters and 957M activations. 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