LongCat-Flash TPS calculator
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 we hold specifications for
Which GPUs can run LongCat-Flash?
Set the inputs, read the answer
A longer conversation needs more memory, which can push this model off smaller cards.
Hides cards that would only fit the model by compressing it below this point.
0 cards match
Calculating| Needs | Quantisation | Fit | |||||
|---|---|---|---|---|---|---|---|
|
No card in our catalogue can run this model with these settings. |
|||||||
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
- Meituan Inc
- Organisation type
- Industry
- Country
- China
- Published
- 1 September 2025
- Authors
- Meituan LongCat Team, Bayan, Bei Li, Bingye Lei, Bo Wang, Bolin Rong, Chao Wang, Chao Zhang, Chen Gao, Chen Zhang, Cheng Sun, Chengcheng Han, Chenguang Xi, Chi Zhang, Chong Peng, Chuan Qin, Chuyu Zhang, Cong Chen, Congkui Wang, Dan Ma, Daoru Pan, Defei Bu, Dengchang Zhao, Deyang Kong, Dishan Liu, Feiye Huo, Fengcun Li, Fubao Zhang, Gan Dong, Gang Liu, Gang Xu, Ge Li, Guoqiang Tan, Guoyuan Lin, Hai…
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, Chat, Code generation, Quantitative reasoning, Instruction interpretation
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
- 560B
- Training data
- 23,000,000,000,000 tokens
"560 billion total parameters, featuring an innovative Mixture-of-Experts (MoE) architecture. The model incorporates a dynamic computation mechanism that activates 18.6B∼31.3B parameters (averaging∼27B)"
"(1) We train the model on approximately 20 trillion tokens with 8192 sequence length to establish a robust base model. (2) Reasoning and coding capabilities are further enhanced using trillions of data. (3) The context length is extended to 128k through training on long context corpora." With "Likely" confidence we may assume theytrained on total of 23T 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
- 3.7 × 10²⁴ FLOP
- How it was established
- Operation counting
6 FLOP/parameter/token * 27000000000 active parameters * 23000000000000 tokens ["Likely" confidence, see dataset size notes] = 3.726e+24 FLOP
The training run
What it physically took to train: which chips, how many, for how long, and what that drew from the wall.
- Wall-clock time
- 720 hours (30 days)
"completing training within 30 days" (=720 hours)
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
- meituan-longcat
MIT license https://huggingface.co/meituan-longcat/LongCat-Flash-Chat
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
Table 3 "LongCat-Flash Base model achieves performance on par with state-of-the-art base models despite its compact active/total parameter size. Although Llama-4-Maverick has fewer activated and total parameters, LongCat-Flash Base surpasses both on nearly all benchmarks." "It achieves the highest score of 89.65 on IFEval, outperforming all other models and demonstrating superior reliability in adhering to complex and nuanced directives. Furthermore, it secures the best score on COLLIE (57.10) …
Sources
Where this record came from and when it was last checked.
- Reference
- LongCat-Flash Technical Report
- Last updated
- 28 November 2025
What the numbers mean
Hardware requirements in practice
At 560B parameters, LongCat-Flash is beyond what any single graphics card holds. Running it means either splitting it across several cards or renting hardware built for the job — 0 of the cards we track can hold it on their own, and all of them are datacentre parts.
About this model
LongCat-Flash was published by Meituan Inc, in China, in September 2025. industry is the category the publisher falls under.
It works in Language, and is recorded as doing language modeling/generation, Question answering, Chat, Code generation, Quantitative reasoning, Instruction interpretation.
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 meituan-longcat organisation on Hugging Face.
Training and provenance
Training it took roughly 3.7 × 10²⁴ FLOP of computation — a measure of what producing the model cost, not of how fast it answers.
It was trained on about 23,000,000,000,000 tokens of text.
The reason it appears in this catalogue at all is sOTA improvement.
Step by step
How to choose a GPU for LongCat-Flash
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 LongCat-Flash. That figure, not the card's headline performance, is what decides whether it runs.
-
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 LongCat-Flash.
-
03
Choose how far you will compress it
The quantisation column varies by card, because a bigger card holds a more accurate copy of LongCat-Flash. Set a floor to hold the comparison at one level.
-
04
Rank by throughput rather than spec sheet
Ranking by tokens per second for LongCat-Flash follows memory bandwidth, not core counts.
-
05
Look at the headroom, not just the fit
Tight means LongCat-Flash 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
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 LongCat-Flash.
Answers
LongCat-Flash — common questions
How accurate are these LongCat-Flash speed estimates?
These are estimates with real error bars. The fastest result here, the range beneath each figure, could reasonably land anywhere in its published range depending on which runtime you use.
Is LongCat-Flash open source?
Its weights are published, so LongCat-Flash 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 LongCat-Flash have?
LongCat-Flash has 560B parameters. "560 billion total parameters, featuring an innovative Mixture-of-Experts (MoE) architecture. The model incorporates a dynamic computation mechanism that activates 18.6B∼31.3B parameters (averaging∼27B)". 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 LongCat-Flash?
LongCat-Flash was published by Meituan Inc, based in China, categorised as industry.
When was LongCat-Flash released?
LongCat-Flash was published in September 2025.
What is LongCat-Flash used for?
LongCat-Flash works in Language, and is recorded as handling language modeling/generation, Question answering, Chat, Code generation, Quantitative reasoning, Instruction interpretation. These are the areas it was designed around; they describe intent rather than a hard boundary.
Where can I download LongCat-Flash?
Its weights are published under the meituan-longcat 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 LongCat-Flash?
Around 3.7 × 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 LongCat-Flash 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 LongCat-Flash is rarely worth using — the nearest miss we calculate is short by 80.3 GB. Every figure here assumes the whole model is on the card.
Would two GPUs run LongCat-Flash faster?
A second card roughly doubles the memory available but not the generation rate. With 0 cards already able to run LongCat-Flash alone, the case for pairing is weak.
Why does the quantisation differ between cards for LongCat-Flash?
Each card is shown running the least-compressed copy it can hold, and LongCat-Flash appears at 1 different compression levels across the cards that fit it. Bigger cards get the more accurate version.
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.