DeepSeekMath-V2 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 DeepSeekMath-V2?
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
- DeepSeek
- Organisation type
- Industry
- Country
- China
- Published
- 27 November 2025
- Authors
- Zhihong Shao, Yuxiang Luo, Chengda Lu, Z. Z. Ren, Jiewen Hu, Tian Ye, Zhibin Gou, Shirong Ma, Xiaokang Zhang
What it does
The problem areas the model was built for. A model can carry several of each.
- Domain
- Language
- Task
- Mathematical reasoning, Theorem proving, Proof generation, Proof verification
- Base model
- DeepSeek-V3.2-Exp-Base
- 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
- 685B
- Training data
- tokens
HF lists 685B params. Built on top of DeepSeek-V3.2-Exp-Base
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
Weights + model card: https://huggingface.co/deepseek-ai/DeepSeek-Math-V2 (Apache-2.0). Inference support referenced to https://github.com/deepseek-ai/DeepSeek-V3.2-Exp.
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
- Confident
Reported gold-level performance on IMO 2025/CMO 2024 and ~118/120 on Putnam 2024 with scaled test-time compute.
Sources
Where this record came from and when it was last checked.
- Reference
- DeepSeekMath-V2: Towards Self-Verifiable Mathematical Reasoning
- Last updated
- 15 January 2026
What the numbers mean
Hardware requirements in practice
DeepSeekMath-V2 reaches a parameter count of 685B. That is beyond what any single graphics card holds. Running it means either splitting it across several cards or renting hardware built for the job, and every card able to hold it alone is a datacentre part. The number that can: 0.
Where it came from
DeepSeekMath-V2 was published by DeepSeek, in the country recorded as China, during November 2025. The category the publisher falls under is industry.
It works in the domain of Language, and is recorded as performing the task of mathematical reasoning, Theorem proving, Proof generation, Proof verification.
Its starting point was an existing base model, DeepSeek-V3.2-Exp-Base. That is the usual way a specialised model is produced.
Because its weights were released, nothing about running it depends on a provider staying available — it is yours once downloaded.
What went into building it
It is tracked in the underlying dataset for one reason in particular: sOTA improvement.
Step by step
How to choose a GPU for DeepSeekMath-V2
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
Every card here has been checked against DeepSeekMath-V2. That figure, not the headline performance of a card, is what decides whether it runs.
-
02
Set the context length you will work at
Set the context to what you will actually use. The cache grows with the conversation, and it is the usual reason a card that seemed fine stops fitting DeepSeekMath-V2.
-
03
Choose how far you will compress it
Each card runs the least-compressed copy it can hold. Setting a minimum quality drops the cards that only manage it by squeezing further than you would want, and holds the comparison at one level.
-
04
Compare tokens per second, not specifications
The speed ordering is effectively an ordering by memory bandwidth, for DeepSeekMath-V2. It will not match a gaming ordering, because generation is bound by memory bandwidth.
-
05
Check the fit verdict before buying
The fit column separates cards that just manage it from those with room to spare, in the case of DeepSeekMath-V2. Comfortable means you can grow the context later. That difference matters more than a few tokens per second, so buy for comfortable if you expect to.
-
06
See what else that card runs
Each card page repeats this sweep for every model we hold, answering what else the hardware is good for beyond DeepSeekMath-V2.
Answers
DeepSeekMath-V2 — common questions
DeepSeekMath-V2— how accurate are these speed estimates?
Every figure is derived from memory bandwidth and model size, not benchmarked, which is why each is published as a range rather than a single number. One example: the range beneath each figure. The same model and card vary by thirty to fifty per cent depending on the inference software and its version.
DeepSeekMath-V2— is it open source?
Its weights are published, so it 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.
DeepSeekMath-V2— how many parameters does it have?
It has a parameter count of 685B. HF lists 685B params. Built on top of DeepSeek-V3.2-Exp-Base. 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.
DeepSeekMath-V2— who created it?
It was published by DeepSeek, based in China, an organisation categorised as industry.
DeepSeekMath-V2— when was it released?
It was published in November 2025.
DeepSeekMath-V2— what is it used for?
It works in the domain of Language, and is recorded as handling the task of mathematical reasoning, Theorem proving, Proof generation, Proof verification. Models frequently carry more than one of each, and the tags describe purpose rather than capability limits.
DeepSeekMath-V2— where can I download it?
The weights are published, though we do not hold a repository link for it. This site calculates hardware requirements rather than hosting model files.
DeepSeekMath-V2— can I run it 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 model is rarely worth using. The nearest miss we calculate falls short by 155.9 GB. Every figure here assumes the whole model is resident on the card.
DeepSeekMath-V2— would two GPUs run it faster?
Capacity adds across cards; throughput does not. The number of cards already holding it on their own: 0. So a second card is rarely the answer here.
DeepSeekMath-V2— why does the quantisation differ between cards?
A larger card holds a more accurate copy. The number of compression levels used across the cards that run it: 1. Bigger cards get the more accurate version, and the quality floor above pins the comparison to one level.
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.