MiniMax-M1-40k 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
Smallest card that fits
Radeon Instinct MI325X
256 GB · Q3_K_M · 11.7 tok/s
Fastest card
B300
18.3 tok/s · 288 GB
Which GPUs can run MiniMax-M1-40k?
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.
4 cards match
Calculating| Needs | Quantisation | Fit | |||||
|---|---|---|---|---|---|---|---|
|
18.3
tok/s
11–29 · low confidence |
B300 NVIDIA | 288 GB | 8,000 GB/s | Sep 2025 | 250.0 GB | IQ4_XS | Tight |
|
14.6
tok/s
9–23 · low confidence |
Radeon Instinct MI350X AMD | 288 GB | 8,190 GB/s | Jan 2025 | 250.0 GB | IQ4_XS | Tight |
|
14.6
tok/s
9–23 · low confidence |
Radeon Instinct MI355X AMD | 288 GB | 8,190 GB/s | Jan 2025 | 250.0 GB | IQ4_XS | Tight |
|
11.7
tok/s
7–19 · low confidence |
Radeon Instinct MI325X AMD | 256 GB | 6,000 GB/s | Oct 2024 | 223.5 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
- MiniMax
- Organisation type
- Industry
- Country
- China
- Published
- 13 June 2025
- Authors
- MiniMax: Aili Chen, Aonian Li, Bangwei Gong, Binyang Jiang, Bo Fei, Bo Yang, Boji Shan, Changqing Yu, Chao Wang, Cheng Zhu, Chengjun Xiao, Chengyu Du, Chi Zhang, Chu Qiao, Chunhao Zhang, Chunhui Du, Congchao Guo, Da Chen, Deming Ding, Dianjun Sun, Dong Li, Enwei Jiao, Haigang Zhou, Haimo Zhang, Han Ding, Haohai Sun, Haoyu Feng, Huaiguang Cai, Haichao Zhu, Jian Sun, Jiaqi Zhuang, Jiaren Cai, Jiayua…
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, Quantitative reasoning
- Base model
- MiniMax-Text-01
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
- 456B
- Training data
- 7,500,000,000,000 tokens
from the base model: Total Parameters: 456B Activated Parameters per Token: 45.9B (matches safetensors)
7.5T tokens "To enhance the reasoning and long context capabilities of the foundation model while ensuring diversity, we continue training the MiniMax-Text-01 model with additional 7.5T tokens with optimized data quality and mixture."
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
- 4.2 × 10²⁴ FLOP
- Fine-tuning compute
- 2.2 × 10²⁴ FLOP
1.9828799999999997e+24 FLOP [base model compute] + 2.2033131e+24 FLOP [see finetune compute notes] = 4.1861931e+24 FLOP
1. Continual pre-training for 7.5T tokens: 6 FLOP / token / active parameter * 45.9 * 10^9 activated parameters * 7.5 * 10^12 tokens = 2.0655e+24 FLOP 2. RL (3 weeks on 512 H800 GPUs for full training, 40k version was stopped in the ~middle of it): 989000000000000 FLOP / GPU / sec [bf16 assumed] * 512 GPUs * 1.5 weeks * 168 hours / week * 3600 sec / hour * 0.3 [assimed utilization] = 1.3781311e+23 FLOP 2.0655e+24 FLOP + 1.3781311e+23 FLOP = 2.2033131e+24 FLOP *there was also SFT stage with …
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
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
- MiniMaxAI
Apache 2,0 https://huggingface.co/MiniMaxAI/MiniMax-M1-40k
How it is classified
Labels the source dataset applies when tracking notable models, and how confident it is in the entry.
- Record confidence
- Likely
Sources
Where this record came from and when it was last checked.
- Reference
- MiniMax-M1: Scaling Test-Time Compute Efficiently with Lightning Attention
- Last updated
- 28 November 2025
The extremes
The ten fastest GPUs for MiniMax-M1-40k
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 MiniMax-M1-40k
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
What you need to run it
Minimum card
Radeon Instinct MI325X
Memory needed
223.5 GB
Fastest
18.3 tok/s
At 456B parameters, MiniMax-M1-40k is beyond what any single graphics card holds. Running it means either splitting it across several cards or renting hardware built for the job — 4 of the cards we track can hold it on their own, and all of them are datacentre parts.
The least hardware that works is a Radeon Instinct MI325X. Its 256 GB is enough at Q3_K_M compression, giving roughly 11.7 tokens per second.
The quickest result comes from a B300 at around 18.3 tokens per second — its 8,000 GB/s of bandwidth is what buys that.
Where it came from
MiniMax-M1-40k was published by MiniMax, in China, in June 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, Quantitative reasoning.
It is derived from MiniMax-Text-01 rather than trained from scratch, which is the usual way a specialised model is produced.
The weights being open is what puts this page in the calculator rather than only in the catalogue: it is a model you can actually hold. It is published under the MiniMaxAI organisation on Hugging Face.
Understanding the speeds
The median result is around 14.6 tokens per second; 4 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.
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
The training run consumed about 4.2 × 10²⁴ FLOP, on NVIDIA H800 SXM5. That figure describes the cost of creating it and has no bearing on how quickly it generates text.
The training set ran to roughly 7,500,000,000,000 tokens.
Step by step
How to choose a GPU for MiniMax-M1-40k
The table above has already assessed every card we hold specifications for against this model. Getting to your answer takes six steps.
-
01
Read the memory figure first
Every card here has been checked against MiniMax-M1-40k — around 223.5 GB at Q3_K_M. Capacity is the gate — a card either holds it or it does not.
-
02
Match the context to your actual use
Set the context to what you will actually use. The cache grows with the conversation, and it is the usual reason MiniMax-M1-40k stops fitting a card that seemed fine.
-
03
Set a quality floor
Each card runs the least-compressed copy it can hold — Q3_K_M on the smallest card that fits. Setting a floor drops the cards that only manage MiniMax-M1-40k by squeezing it further than you would want.
-
04
Sort by speed
The speed ordering for MiniMax-M1-40k is effectively an ordering by memory bandwidth, which is why the B300 tops it at 18.3 tok/s.
-
05
Read the fit column last
The fit column separates cards that just manage MiniMax-M1-40k from those with room to spare. Buy for the second if the context might grow.
-
06
Open the card you have settled on
Every card name links to its own page, which runs the same calculation across the whole model catalogue. Worth a look before buying for MiniMax-M1-40k alone — a card is usually bought for more than one model.
Answers
MiniMax-M1-40k — common questions
How much VRAM does MiniMax-M1-40k need?
About 223.5 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.
Is MiniMax-M1-40k open source?
Its weights are published, so MiniMax-M1-40k 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 MiniMax-M1-40k have?
MiniMax-M1-40k has 456B parameters. from the base model: Total Parameters: 456B Activated Parameters per Token: 45.9B (matches safetensors). 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 MiniMax-M1-40k?
MiniMax-M1-40k was published by MiniMax, based in China, categorised as industry.
When was MiniMax-M1-40k released?
MiniMax-M1-40k was published in June 2025.
What is MiniMax-M1-40k used for?
MiniMax-M1-40k works in Language, and is recorded as handling language modeling/generation, Question answering, Code generation, Quantitative reasoning. These are the areas it was designed around; they describe intent rather than a hard boundary.
Where can I download MiniMax-M1-40k?
Its weights are published under the MiniMaxAI 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 MiniMax-M1-40k?
Around 4.2 × 10²⁴ 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 MiniMax-M1-40k if it does not fit in my GPU?
It can be split between the card and system memory, but MiniMax-M1-40k generates painfully slowly that way — the nearest miss we calculate is short by 103.8 GB. Nothing on this page assumes offloading.
Would two GPUs run MiniMax-M1-40k faster?
A second card roughly doubles the memory available but not the generation rate. With 4 cards already able to run MiniMax-M1-40k alone, the case for pairing is weak.
Why does the quantisation differ between cards for MiniMax-M1-40k?
Each card is shown running the least-compressed copy it can hold, and MiniMax-M1-40k appears at 2 different compression levels across the cards that fit it. Bigger cards get the more accurate version.
How accurate are these MiniMax-M1-40k 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 11–29 tok/s on the B300 rather than a single number.
What GPU do I need to run MiniMax-M1-40k?
The smallest card in our catalogue that holds MiniMax-M1-40k is the Radeon Instinct MI325X, with 256 GB of memory. It runs the model at Q3_K_M using about 223.5 GB, and produces roughly 11.7 tokens per second. 4 cards in total can run it.
How fast is MiniMax-M1-40k on a GPU?
It depends on the card. The quickest we calculate is a B300 at about 18.3 tokens per second; the slowest that still runs it manages considerably less. Reading speed is around ten tokens per second, and 4 of the cards that can run MiniMax-M1-40k clear that.
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.