Calculate the TPS of the GRID M3-3020 on local AI models

NVIDIA 4 GB GDDR5 83 GB/s May 2016

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

97 models it can run

679 models in our catalogue altogether

Largest model it holds

DeciLM 6B

5.7B · Q3_K_M · 14.2 tok/s

Fastest model

Gemma 3 QAT 1B

30.0 tok/s · 1B

Which AI models can run on a GRID M3-3020?

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.

97 models match

Calculating
Quantisation Fit
30.0 tok/s

10–60 · low confidence

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

10–60 · low confidence

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

10–60 · low confidence

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

10–60 · low confidence

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

10–60 · low confidence

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

10–60 · low confidence

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

10–55 · low confidence

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

10–54 · low confidence

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

10–54 · low confidence

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

10–54 · low confidence

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

10–54 · low confidence

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

9–50 · low confidence

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

9–50 · low confidence

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

9–50 · low confidence

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

9–50 · low confidence

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

9–49 · low confidence

Llama 3.2 1B 1.2B Sep 2024 2.2 GB 54k tokens Q8_0 Comfortable
24.0 tok/s

8–48 · low confidence

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

8–46 · low confidence

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

8–46 · low confidence

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

8–46 · low confidence

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

8–46 · low confidence

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

8–46 · low confidence

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

8–46 · low confidence

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

8–46 · low confidence

Otter 1.3B May 2023 2.1 GB 131k tokens ? Q8_0 Comfortable
23.0 tok/s

8–46 · low confidence

Phi-1 1.3B Oct 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 M3-3020 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
4 GB
Memory bandwidth
83 GB/s
Memory type
GDDR5
Memory bus width
128 bit
Memory clock
1.3 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
GM107
Architecture
Maxwell
Generation
GRID(Mx)
Foundry
TSMC
Process size
28 nm
Transistors
1.9 billion
Transistor density
12,600 K/mm²
Die size
148 mm²
Package
FCBGA-908
Released
18 May 2016

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
1.03 GHz
Boost clock
1.31 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
640
Texture mapping units
40
Render output units
16
Streaming multiprocessors
5
L1 cache
64 KB
L2 cache
2 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.

Single precision (FP32)
1.7 TFLOPS
Double precision (FP64)
52.2 GFLOPS
Pixel rate
21 GPixel/s
Texture rate
52 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.

Suggested power supply
200 W
Power connectors
1x 8-pin
Bus interface
PCIe 3.0 x16
Slot width
Dual-slot
Dimensions
267 mm

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
5.0
DirectX
11.0
OpenGL
4.6
Vulkan
1.4
OpenCL
3.0
Shader model
5.1

Listings

Where to buy a GRID M3-3020

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

Why memory is the number that matters here

Memory

4 GB

Bandwidth

83 GB/s

Largest model

DeciLM 6B

At 4 GB of GDDR5 the GRID M3-3020 is limited to the smaller end of the catalogue. About 3.6 GB is actually available to a runtime, and a model has to fit entirely inside it before generating anything at all.

At 83 GB/s across a 128-bit bus, bandwidth is this card's real constraint. Every token requires reading the entire model out of memory, so a large model will feel slow here even when it fits.

That comes from a 1.3 GHz memory clock across the bus width above. 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 DeciLM 6B — 5.7B, compressed to Q3_K_M, generating around 14.2 tokens per second.

The chip and how it was built

The GRID M3-3020 is built on the GM107 graphics processor, using NVIDIA's Maxwell architecture, as part of the GRID(Mx) generation.

The chip is manufactured by TSMC, on a 28 nm process, with a die measuring 148 mm², holding 1.9 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 2016, roughly 10 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

FP64

52.2 GFLOPS

Double-precision throughput is 52.2 GFLOPS. 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.

Clocks run from 1.03 GHz at base to 1.31 GHz boosted. 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

The GRID M3-3020 has 64 KB of L1 cache, backed by 2 MB of L2. 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 640 shading units, 40 texture mapping units, and 16 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

The board occupies a dual-slot, measuring 267 mm long, and needs 1x 8-pin. 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 3.0 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 M3-3020

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 Qwen3.5-4B 4B · Q5_K_M · Feb 2026 13.4 tok/s
  2. 02 Voxtral Mini 4.7B · Q4_K_M · Jul 2025 14.7 tok/s
  3. 03 Typhoon 2.1 Gemma 4B 4B · Q4_K_M · May 2025 17.3 tok/s
  4. 04 Qwen3-4B 4B · Q3_K_M · Apr 2025 20.2 tok/s
  5. 05 Gemma 3 QAT 4B 4B · Q4_K_M · Apr 2025 17.3 tok/s
  6. 06 Gemma 3 4B 4B · Q4_K_M · Mar 2025 17.3 tok/s
  7. 07 Phi-4-Multimodal 5.6B · Q3_K_M · Mar 2025 14.4 tok/s
  8. 08 Minitron 4B 4.2B · Q4_K_M · Nov 2024 16.5 tok/s
  9. 09 XVERSE-MoE-A4.2B 4.2B · Q4_K_M · Apr 2024 16.5 tok/s
  10. 10 DeciLM 6B 5.7B · Q3_K_M · Sep 2023 14.2 tok/s

The fastest AI models on a GRID M3-3020

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 30.0 tok/s
  2. 02 Gemma 3 1B 1B · Q8_0 · 1.8 GB 30.0 tok/s
  3. 03 LLama 3..2 Typhoon 2 1B 1B · Q8_0 · 1.8 GB 30.0 tok/s
  4. 04 OLMo-1B 1B · Q8_0 · 1.8 GB 30.0 tok/s
  5. 05 HGRN 1B (WT 103) 1B · Q8_0 · 1.8 GB 30.0 tok/s
  6. 06 Pythia-1b 1B · Q8_0 · 1.8 GB 30.0 tok/s
  7. 07 OpenELM-1.1B 1.1B · Q8_0 · 1.9 GB 27.7 tok/s
  8. 08 TinyLlama-1.1B (1T token checkpoint) 1.1B · Q8_0 · 1.9 GB 27.2 tok/s
  9. 09 TinyLlama-1.1B (3T token checkpoint) 1.1B · Q8_0 · 1.9 GB 27.2 tok/s
  10. 10 DeciCoder-1B 1.1B · Q8_0 · 1.9 GB 27.2 tok/s

Step by step

How to work out the tokens per second of a GRID M3-3020

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 97 models this GRID M3-3020 can run. Search by name, or by size — typing 27b matches on the parameter count even when the name never states it.

  2. 02

    Set the context length you will actually use

    Longer conversations cost memory on top of the weights. With 4 GB to work in, that is frequently the difference between a model fitting and not.

  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

    Speeds come with error bars for a reason. The best case here is 30.0 tok/s on Gemma 3 QAT 1B, and which inference software you use moves that by thirty to fifty per cent.

  5. 05

    Read the fit verdict last

    The fit column separates models that just fit from those with room to spare — worth checking against the card's 4 GB before settling on one.

  6. 06

    Cross-check against other hardware

    Each model page repeats this calculation for the whole catalogue. Worth a look before deciding: it shows what else runs the same model, and how the GRID M3-3020 compares.

Answers

GRID M3-3020 — common questions

01

What is the largest AI model a GRID M3-3020 can run?

The largest model in our catalogue that fits on a GRID M3-3020 is DeciLM 6B at 5.7B parameters, compressed to Q3_K_M. It generates roughly 14.2 tokens per second and needs about 3.5 GB of the card's memory.

02

How many tokens per second does a GRID M3-3020 produce?

It depends on the model. On a GRID M3-3020 the fastest model we track is Gemma 3 QAT 1B at about 30.0 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.

03

How much memory does a GRID M3-3020 have?

A GRID M3-3020 has 4 GB of GDDR5 memory. Around a tenth of that is reserved by the inference runtime and the driver, leaving roughly 3.6 GB available for a model and its conversation.

04

What is the memory bandwidth of a GRID M3-3020?

The GRID M3-3020 has 83 GB/s of memory bandwidth, across a 128-bit memory bus. 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.

05

What type of memory does a GRID M3-3020 use?

It uses GDDR5 clocked at 1.3 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.

06

Who makes the GRID M3-3020?

The GRID M3-3020 is a NVIDIA product, with the chip manufactured by TSMC, on a 28 nm process.

07

When was the GRID M3-3020 released?

The GRID M3-3020 was released in May 2016.

08

How much cache does a GRID M3-3020 have?

The GRID M3-3020 has 64 KB of L1 cache, and 2 MB of L2 cache. 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.

09

Does the GRID M3-3020 support CUDA?

Yes. The GRID M3-3020 reports CUDA compute capability 5.0, which predates tensor cores. 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.

10

What bus interface does the GRID M3-3020 use?

It uses PCIe 3.0 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.

11

Is the GRID M3-3020 good for running local AI models?

Its memory limits it to smaller models though its bandwidth means generation will feel slow on larger models. In total it runs 97 of the models we track. Whether that is enough depends entirely on which model you want — the table above answers that directly.

12

Can a GRID M3-3020 run a model that does not fit in its memory?

Only partly. Layers beyond the 4 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.

13

Would two GRID M3-3020 cards be twice as fast?

No. A second GRID M3-3020 doubles the memory to 8 GB, which lets you hold models neither could hold alone, but generation does not split that way. These figures describe one card.

14

What AI models can a GRID M3-3020 run?

97 of the 679 open-weight language models we track fit on a GRID M3-3020 and can be run locally on it. The table on this page lists every one, with the memory it needs, the quantisation it runs at and an estimated generation speed.

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