Calculate the TPS of the GRID A100B on local AI models

NVIDIA 48 GB HBM2e 1,870 GB/s May 2020

Every model in our catalogue assessed against this card at the context length and minimum quality you choose. Speed is an estimate for a single request, calculated from this card's memory bandwidth and the size of each model once compressed.

Calculated for this card

607 models it can run

721 models in our catalogue altogether

Largest model it holds

Qwen3-Coder-Next

80B · IQ4_XS · 135 tok/s

Fastest model

Gemma 3 QAT 1B

792 tok/s · 1B

Which AI models can run on a GRID A100B?

Set the inputs, read the answer

More context means more memory for the conversation cache. Speed is for a fresh conversation and does not change with this setting.

Hides models that would only fit by being compressed below this point.

607 models match

Calculating
Quantisation Fit
792 tok/s

673–950

Gemma 3 1B 1B Mar 2025 1.8 GB 33k tokens Q8_0 Comfortable
792 tok/s

673–950

Gemma 3 QAT 1B 1B Apr 2025 1.8 GB 33k tokens Q8_0 Comfortable
792 tok/s

475–1,267 · low confidence

HGRN 1B (WT 103) 1B Nov 2023 1.8 GB 131k tokens ? Q8_0 Comfortable
792 tok/s

475–1,267 · low confidence

LLama 3..2 Typhoon 2 1B 1B Dec 2024 1.8 GB 131k tokens ? Q8_0 Comfortable
792 tok/s

475–1,267 · low confidence

OLMo-1B 1B Feb 2024 1.8 GB 131k tokens ? Q8_0 Comfortable
792 tok/s

475–1,267 · low confidence

Pythia-1b 1B Apr 2023 1.8 GB 131k tokens ? Q8_0 Comfortable
733 tok/s

440–1,173 · low confidence

OpenELM-1.1B 1.1B May 2024 1.9 GB 131k tokens ? Q8_0 Comfortable
720 tok/s

432–1,152 · low confidence

DeciCoder-1B 1.1B Aug 2023 1.9 GB 131k tokens ? Q8_0 Comfortable
720 tok/s

432–1,152 · low confidence

SantaCoder 1.1B Jan 2023 1.9 GB 131k tokens ? Q8_0 Comfortable
720 tok/s

432–1,152 · low confidence

TinyLlama-1.1B (1T token checkpoint) 1.1B Oct 2023 1.9 GB 131k tokens ? Q8_0 Comfortable
720 tok/s

432–1,152 · low confidence

TinyLlama-1.1B (3T token checkpoint) 1.1B Oct 2023 1.9 GB 131k tokens ? Q8_0 Comfortable
660 tok/s

396–1,056 · low confidence

EXAONE 4.0 (1.2B) 1.2B Jul 2025 2.0 GB 131k tokens ? Q8_0 Comfortable
660 tok/s

396–1,056 · low confidence

LFM2-1.2B 1.2B Jul 2025 2.0 GB 131k tokens ? Q8_0 Comfortable
660 tok/s

396–1,056 · low confidence

MinerU2.5 1.2B Sep 2025 2.0 GB 131k tokens ? Q8_0 Comfortable
660 tok/s

396–1,056 · low confidence

Pleias 1.0 1.2B 1.2B Dec 2024 2.0 GB 131k tokens ? Q8_0 Comfortable
660 tok/s

396–1,056 · low confidence

Pleias-RAG-1B 1.2B Apr 2025 2.0 GB 131k tokens ? Q8_0 Comfortable
644 tok/s

547–773

Llama 3.2 1B 1.2B Sep 2024 2.2 GB 131k tokens Q8_0 Comfortable
635 tok/s

381–1,016 · low confidence

MiniCPM-1.2B 1.2B Jun 2024 2.0 GB 131k tokens ? Q8_0 Comfortable
609 tok/s

366–975 · low confidence

DeepSeek Coder 1.3B 1.3B Jan 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
609 tok/s

366–975 · low confidence

DeepSeek-VL-1.3B 1.3B Mar 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
609 tok/s

366–975 · low confidence

DigiRL 1.3B Jun 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
609 tok/s

366–975 · low confidence

GLA Transformer 1.3B 1.3B Aug 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
609 tok/s

366–975 · low confidence

Janus 1.3B 1.3B Oct 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
609 tok/s

366–975 · low confidence

Kosmos-2.5 1.3B Aug 2024 2.1 GB 131k tokens ? Q8_0 Comfortable
609 tok/s

366–975 · low confidence

Otter 1.3B May 2023 2.1 GB 131k tokens ? Q8_0 Comfortable

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

GRID A100B full specification

Everything on record for this board, ordered by how much it bears on running a language model rather than by how a spec sheet would list it. Memory comes first because it decides the outcome; the rest is context.

Memory

The two specifications that decide what this card can run and how quickly. Capacity sets which models fit; bandwidth sets how many tokens per second they produce once they do.

Memory size
48 GB
Memory bandwidth
1,870 GB/s
Memory type
HBM2e
Memory bus width
6,144 bit
Memory clock
1.22 GHz

The chip

Which processor is on the board and how it was manufactured. A smaller process size generally means more performance for the same power.

Graphics processor
GA100
Architecture
Ampere
Generation
GRID(Ax)
Foundry
TSMC
Process size
7 nm
Transistors
54.2 billion
Transistor density
65,600 K/mm²
Die size
826 mm²
Package
BGA-2743
Released
14 May 2020

Clock speeds

How fast the processor runs. Worth far less here than on a gaming benchmark: generating text is limited by memory bandwidth, so a higher clock barely moves the result.

Base clock
900 MHz
Boost clock
1.01 GHz

Processing units

What the chip contains. These drive graphics performance and matter mainly for processing a long prompt rather than for producing the answer.

Shading units
6,912
Texture mapping units
432
Render output units
192
Streaming multiprocessors
108
Tensor cores
432
L1 cache
192 KB
L2 cache
48 MB

Theoretical performance

Peak arithmetic rates published for the board. These are ceilings that no real workload reaches, and generating text reaches a small fraction of them because it is limited by memory rather than arithmetic.

Half precision (FP16)
55.6 TFLOPS
Single precision (FP32)
13.9 TFLOPS
Double precision (FP64)
6.9 TFLOPS
Pixel rate
193 GPixel/s
Texture rate
434 GTexel/s

The board

What it takes to physically install and power the card — the practical constraints that decide whether it fits the machine you already own.

Power draw (TDP)
400 W
Suggested power supply
800 W
Power connectors
None
Bus interface
PCIe 4.0 x16
Slot width
IGP

Software support

Which graphics and compute interfaces the card supports. CUDA compute capability is the one that bears on inference: below 7.0 there are no tensor cores, and modern inference software falls back to slower code paths.

CUDA compute capability
8.0
OpenCL
3.0

Listings

Where to buy a GRID A100B

No vendor is currently listing this card. Listings come from vendors who publish them here directly — browse the vendor directory to see who is selling what.

What the numbers mean

Capacity and bandwidth

Memory

48 GB

Bandwidth

1,870 GB/s

Largest model

Qwen3-Coder-Next

GRID A100B carries 48 GB of HBM2e. That covers the mid-sized models most people actually run. Once the runtime and driver reserve their working space, roughly this much is left: 43.2 GB.

Memory bandwidth reaches 1,870 GB/s across a bus of 6,144 bits. Generating a token means reading every weight once, so that figure sets the pace more than any other number here, and at this level text arrives faster than most people read.

That comes from a memory clock of 1.22 GHz. Widening the bus and raising the clock are the two levers a manufacturer has, which is why a card with unremarkable cores can still generate quickly.

In practice that combination tops out at Qwen3-Coder-Next, 80B, compressed to IQ4_XS and generating around 135 tokens per second.

The chip and how it was built

GRID A100B is built on the graphics processor GA100, using the architecture Ampere from NVIDIA, as part of the generation GRID(Ax).

The chip is manufactured by TSMC, on a process of 7 nm, with a die measuring 826 mm², holding 54.2 billion transistors. A smaller process generally means more performance for the same power, though for language models it matters far less than the memory subsystem.

It was released in May 2020, roughly 6.3354499963488 years ago. Inference software support tends to follow hardware by a year or two, so a card of this age generally has mature, well-optimised code paths available to it.

Compute throughput, and why it matters less than it looks

FP16

55.6 TFLOPS

FP64

6.9 TFLOPS

Tensor cores

432

On paper GRID A100B reaches 55.6 TFLOPS at half precision, and 13.9 TFLOPS at single precision. These are peak figures no real workload sustains, and generating text reaches only a small fraction of them — decoding is limited by memory rather than arithmetic, which is why a card can look enormously powerful here and still produce tokens at an ordinary rate.

Double-precision throughput reaches 6.9 TFLOPS. It has no bearing on running a language model — no inference runtime uses it — but it separates datacentre parts from consumer ones, since the latter deliberately restrict it.

The card carries 432 tensor cores across 108 streaming multiprocessors. These accelerate the matrix arithmetic at the heart of a transformer, and they are what make prompt processing — reading a long document before answering — dramatically faster than it would otherwise be.

Clocks run from a base of 900 MHz to a boost of 1.01 GHz. Worth far less here than on a gaming benchmark: raising the clock speeds up the arithmetic, and the arithmetic is not what generation is waiting on.

Cache and processing units

GRID A100B has an L1 cache of 192 KB, backed by an L2 cache of 48 MB. Cache absorbs a share of the memory traffic that would otherwise hit the main bus, which is the one place on this page where a number other than bandwidth quietly affects generation speed — a large L2 lets more of the working set stay close to the cores.

There are 6,912 shading units, 432 texture mapping units, and 192 render output units. These drive graphics workloads and contribute to prompt processing, but they sit idle for much of the time a model spends generating a reply.

Power, size and installation

Power draw

400 W

GRID A100B is rated at 400 W, and the suggested system power supply is 800 W. Running a language model keeps a card busy in bursts rather than continuously — it draws hard while generating and idles between requests — so sustained draw over a working day is usually well below the rated figure.

The board occupies igp. Worth checking against the case and power supply already in the machine, since the largest cards need considerably more of both than a typical desktop provides.

It connects over PCIe 4.0 x16. The interface governs how quickly a model is loaded from disk into the card, not how fast it runs once there, so a narrower link costs a few seconds at startup and nothing thereafter.

The extremes

The largest AI models that run on a GRID A100B

The biggest open-weight models that fit on this card, newest first. Each is shown at the best compression the card can hold.

  1. 01 Qwen3-Coder-Next 80B · IQ4_XS · Feb 2026 135 tok/s
  2. 02 Qwen3-Next-80B-A3B 80B · IQ4_XS · Sep 2025 135 tok/s
  3. 03 Kimi Dev 72b 72B · IQ4_XS · Jun 2025 27.0 tok/s
  4. 04 OpenThaiGPT 1.6 / OTG-1.6 (72B) 72B · IQ4_XS · Apr 2025 27.0 tok/s
  5. 05 InternVL2_5-78B 78.4B · Q3_K_M · Dec 2024 27.3 tok/s
  6. 06 Qwen2.5-72B 72.7B · Q4_K_M · Sep 2024 25.2 tok/s
  7. 07 Qwen2.5 Instruct (72B) 72.7B · Q4_K_M · Sep 2024 25.2 tok/s
  8. 08 InternVL2-Llama3-76B 76B · IQ4_XS · Jul 2024 25.6 tok/s
  9. 09 Qwen2-72B 72.7B · Q4_K_M · Jun 2024 25.2 tok/s
  10. 10 IDEFICS-80B 80B · Q3_K_M · Aug 2023 26.7 tok/s

The fastest AI models on a GRID A100B

Where this card produces tokens quickest. Smaller models dominate here, because generating each token means reading the whole model out of memory once.

  1. 01 Gemma 3 QAT 1B 1B · Q8_0 · 1.8 GB 792 tok/s
  2. 02 Gemma 3 1B 1B · Q8_0 · 1.8 GB 792 tok/s
  3. 03 LLama 3..2 Typhoon 2 1B 1B · Q8_0 · 1.8 GB 792 tok/s
  4. 04 OLMo-1B 1B · Q8_0 · 1.8 GB 792 tok/s
  5. 05 HGRN 1B (WT 103) 1B · Q8_0 · 1.8 GB 792 tok/s
  6. 06 Pythia-1b 1B · Q8_0 · 1.8 GB 792 tok/s
  7. 07 OpenELM-1.1B 1.1B · Q8_0 · 1.9 GB 733 tok/s
  8. 08 TinyLlama-1.1B (1T token checkpoint) 1.1B · Q8_0 · 1.9 GB 720 tok/s
  9. 09 TinyLlama-1.1B (3T token checkpoint) 1.1B · Q8_0 · 1.9 GB 720 tok/s
  10. 10 DeciCoder-1B 1.1B · Q8_0 · 1.9 GB 720 tok/s

Step by step

How to work out the tokens per second of a GRID A100B

You do not have to calculate anything by hand — the gputps.com calculator on this page has already worked it out for every model this card can hold. Reading off the answer takes six steps.

  1. 01

    Search for the model you want

    The table lists 607 models this card can run. Search by name, or by size — typing 27b matches on the parameter count even when the name never states it.

  2. 02

    Decide how long your conversations run

    Set the context to your real working length. Short questions cost almost nothing, but a long document can consume a large share of 48 GB so the setting is worth getting right.

  3. 03

    Choose how far you will compress

    Each model is shown at the best compression this card can hold. A minimum quality hides the ones that only fit by being squeezed further than you would accept.

  4. 04

    Read the speed and the range

    Each speed is an estimate for a single conversation, with a range beneath it. The top end here is 792 tok/s on Gemma 3 QAT 1B. The same card and model vary by thirty to fifty per cent between inference runtimes.

  5. 05

    Read the fit verdict last

    The fit column separates models that just fit from those with room to spare — worth checking before settling on one, against an available 48 GB.

  6. 06

    Check the same model from the other side

    Each model page repeats this calculation for the whole catalogue. Worth a look before deciding: it shows what else runs the same model, alongside GRID A100B.

Answers

GRID A100B — common questions

01

GRID A100B— how many tensor cores does it have?

It has 432 tensor cores across 108 streaming multiprocessors. They accelerate the matrix arithmetic a transformer is built from, which mainly speeds up processing a long prompt rather than producing the reply.

02

GRID A100B— does it support CUDA?

Yes. It reports CUDA compute capability 8.0. Capability 7.0 and above has tensor cores, which modern inference software uses; below that it falls back to slower code paths for quantised models.

03

GRID A100B— what bus interface does it use?

It uses PCIe 4.0 x16. This governs how fast a model is loaded onto the card rather than how fast it runs once loaded, so it costs a few seconds at startup and nothing during generation.

04

GRID A100B— is it good for running local AI models?

Its memory is large enough for models most desktop hardware cannot touch and its bandwidth is high enough to generate text faster than most people read. In total it runs 607 of the models we track. Whether that is enough depends entirely on which model you want — the table above answers that directly.

05

GRID A100B— can it run a model that does not fit in its memory?

It can be split, with the overflow held in system memory beyond the card's 48 GB sit in system memory and run at a fraction of the speed, so a mostly-offloaded model is rarely worth using. Every figure here assumes it is fully resident on the card.

06

Would two GRID A100B cards be twice as fast?

Pairing them buys headroom rather than pace: 96 GB of combined memory, at roughly the same generation speed as one.

07

GRID A100B— which AI models can it run?

607 of the 721 open-weight language models we track fit on this card and can be run locally. The table on this page lists every one, with the memory it needs, the quantisation it runs at and an estimated generation speed.

08

GRID A100B— what is the largest AI model it can run?

The largest model in our catalogue that fits is Qwen3-Coder-Next at 80B parameters, compressed to IQ4_XS. It generates roughly 135 tokens per second and needs about 38.8 GB of the card's memory.

09

GRID A100B— how many tokens per second does it produce?

It depends on the model. The fastest model we track here is Gemma 3 QAT 1B at about 792 tokens per second, while larger models run proportionally slower because each token requires reading the whole model out of memory once. Speeds are estimates for a single conversation at a time.

10

GRID A100B— can it run 7B models?

Yes. For example it runs Gemma 4 E4B at Q8_0, using about 9.8 GB of memory and generating around 176 tokens per second.

11

GRID A100B— can it run 13B models?

Yes. For example it runs DeepSeekMoE-16B at Q8_0, using about 17.5 GB of memory and generating around 275 tokens per second.

12

GRID A100B— can it run 30B models?

Yes. For example it runs ERNIE-4.5-VL-28B-A3B at Q8_0, using about 29.2 GB of memory and generating around 157 tokens per second.

13

GRID A100B— can it run 70B models?

Yes. For example it runs Qwen3-Coder-Next at IQ4_XS, using about 38.8 GB of memory and generating around 135 tokens per second.

14

GRID A100B— how much memory does it have?

This card has 48 GB of HBM2e. Around a tenth is reserved by the inference runtime and the driver, leaving roughly 43.2 GB available for a model and its conversation.

15

GRID A100B— what is its memory bandwidth?

Memory bandwidth reaches 1,870 GB/s across a bus of 6,144 bits. This is the single best predictor of how fast it generates text, because producing each token means reading the entire model out of memory once.

16

GRID A100B— what type of memory does it use?

It uses HBM2e clocked at 1.22 GHz. HBM types are found on datacentre accelerators and carry far more bandwidth than the GDDR used on desktop cards, which is why they generate tokens considerably faster at the same capacity.

17

GRID A100B— who makes it?

This is a product of NVIDIA, with the chip manufactured by TSMC, on a process of 7 nm.

18

GRID A100B— when was it released?

It was released in May 2020.

19

GRID A100B— how much power does it use?

Rated board power is 400 W, and the suggested system power supply is 800 W. Generating text draws hard in bursts and idles between requests, so average consumption over a working session is normally well below the rated figure.

20

GRID A100B— how much cache does it have?

The L1 cache is 192 KB, and the L2 cache is 48 MB. Cache absorbs part of the memory traffic that would otherwise reach the main bus, so a larger L2 gives a modest lift to generation speed beyond what bandwidth alone predicts.

21

GRID A100B— what are its TFLOPS?

It is rated at 55.6 TFLOPS at half precision and 13.9 TFLOPS at single precision. These are peak arithmetic ceilings rather than achievable rates, and text generation reaches only a small fraction of them because it is limited by memory bandwidth instead.

The other direction

Looking at it from the other side?

This page starts from the hardware. If you already know which model you want and need to know what it takes to run it, start from the model instead.

All GPUs