Calculate the TPS of the GeForce GTX 960 OEM on local AI models

NVIDIA 4 GB GDDR5 112 GB/s November 2015

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 · 19.1 tok/s

Fastest model

Gemma 3 QAT 1B

40.4 tok/s · 1B

Which AI models can run on a GeForce GTX 960 OEM?

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
40.4 tok/s

14–81 · low confidence

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

14–81 · low confidence

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

14–81 · low confidence

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

14–81 · low confidence

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

14–81 · low confidence

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

14–81 · low confidence

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

13–75 · low confidence

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

13–73 · low confidence

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

13–73 · low confidence

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

13–73 · low confidence

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

13–73 · low confidence

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

12–67 · low confidence

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

12–67 · low confidence

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

12–67 · low confidence

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

12–67 · low confidence

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

11–66 · low confidence

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

11–65 · low confidence

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

11–62 · low confidence

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

11–62 · low confidence

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

11–62 · low confidence

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

11–62 · low confidence

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

11–62 · low confidence

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

11–62 · low confidence

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

11–62 · low confidence

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

11–62 · 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

GeForce GTX 960 OEM 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
112 GB/s
Memory type
GDDR5
Memory bus width
128 bit
Memory clock
1.75 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
GM206
Architecture
Maxwell 2.0
Generation
GeForce 900
Foundry
TSMC
Process size
28 nm
Transistors
2.9 billion
Transistor density
12,900 K/mm²
Die size
228 mm²
Package
BGA-1428
Released
26 November 2015

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
1.18 GHz
Boost clock
1.2 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
1,024
Texture mapping units
64
Render output units
32
Streaming multiprocessors
8
L1 cache
48 KB
L2 cache
1 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)
2.5 TFLOPS
Double precision (FP64)
76.9 GFLOPS
Pixel rate
38 GPixel/s
Texture rate
77 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
None
Bus interface
PCIe 3.0 x16
Slot width
Dual-slot

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
5.2
DirectX
12.1
OpenGL
4.6
Vulkan
1.4
OpenCL
3.0
Shader model
6.8

Listings

Where to buy a GeForce GTX 960 OEM

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

What the memory subsystem means for AI

Memory

4 GB

Bandwidth

112 GB/s

Largest model

DeciLM 6B

GeForce GTX 960 OEM carries only 4 GB of GDDR5. That limits it to the smaller end of the catalogue, and a model has to fit entirely inside before it generates anything at all. A runtime actually gets about 3.6 GB.

Memory bandwidth reaches 112 GB/s across a bus of 128 bits. 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 memory clock of 1.75 GHz. It is why core counts predict generation speed so poorly.

The practical ceiling is DeciLM 6B, 5.7B, compressed to Q3_K_M and generating around 19.1 tokens per second.

The chip and how it was built

GeForce GTX 960 OEM is built on the graphics processor GM206, using the architecture Maxwell 2.0 from NVIDIA, as part of the generation GeForce 900.

The chip is manufactured by TSMC, on a process of 28 nm, with a die measuring 228 mm², holding 2.9 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 November 2015, roughly 10.679476546987 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

76.9 GFLOPS

Double-precision throughput reaches 76.9 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 a base of 1.18 GHz to a boost of 1.2 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

GeForce GTX 960 OEM has an L1 cache of 48 KB, backed by an L2 cache of 1 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 1,024 shading units, 64 texture mapping units, and 32 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 dual-slot. 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 GeForce GTX 960 OEM

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 18.0 tok/s
  2. 02 Voxtral Mini 4.7B · Q4_K_M · Jul 2025 19.8 tok/s
  3. 03 Typhoon 2.1 Gemma 4B 4B · Q4_K_M · May 2025 23.3 tok/s
  4. 04 Qwen3-4B 4B · Q3_K_M · Apr 2025 27.3 tok/s
  5. 05 Gemma 3 QAT 4B 4B · Q4_K_M · Apr 2025 23.3 tok/s
  6. 06 Gemma 3 4B 4B · Q4_K_M · Mar 2025 23.3 tok/s
  7. 07 Phi-4-Multimodal 5.6B · Q3_K_M · Mar 2025 19.5 tok/s
  8. 08 Minitron 4B 4.2B · Q4_K_M · Nov 2024 22.2 tok/s
  9. 09 XVERSE-MoE-A4.2B 4.2B · Q4_K_M · Apr 2024 22.2 tok/s
  10. 10 DeciLM 6B 5.7B · Q3_K_M · Sep 2023 19.1 tok/s

The fastest AI models on a GeForce GTX 960 OEM

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 40.4 tok/s
  2. 02 Gemma 3 1B 1B · Q8_0 · 1.8 GB 40.4 tok/s
  3. 03 LLama 3..2 Typhoon 2 1B 1B · Q8_0 · 1.8 GB 40.4 tok/s
  4. 04 OLMo-1B 1B · Q8_0 · 1.8 GB 40.4 tok/s
  5. 05 HGRN 1B (WT 103) 1B · Q8_0 · 1.8 GB 40.4 tok/s
  6. 06 Pythia-1b 1B · Q8_0 · 1.8 GB 40.4 tok/s
  7. 07 OpenELM-1.1B 1.1B · Q8_0 · 1.9 GB 37.4 tok/s
  8. 08 TinyLlama-1.1B (1T token checkpoint) 1.1B · Q8_0 · 1.9 GB 36.7 tok/s
  9. 09 TinyLlama-1.1B (3T token checkpoint) 1.1B · Q8_0 · 1.9 GB 36.7 tok/s
  10. 10 DeciCoder-1B 1.1B · Q8_0 · 1.9 GB 36.7 tok/s

Step by step

How to work out the tokens per second of a GeForce GTX 960 OEM

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

    Start with the model, not the specification

    The table lists 97 models this card runs. Search narrows the list by name or by size.

  2. 02

    Match the context to your work

    Longer conversations cost memory on top of the weights. Against 4 GB it is often what pushes a large model over the edge.

  3. 03

    Choose how far you will compress

    By default the table picks the least-compressed copy that fits. Setting a floor removes models that only qualify through heavy compression.

  4. 04

    Look at the range, not just the number

    The figures are calculated, not measured. The fastest result on this card is 40.4 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

    Tight means it works today; comfortable means it still works when the conversation grows. Compare what each model needs against an available 4 GB.

  6. 06

    Open the model to compare cards

    Each model page repeats this calculation for the whole catalogue. Worth a look before deciding: it shows what else runs the same model, alongside GeForce GTX 960 OEM.

Answers

GeForce GTX 960 OEM — common questions

01

GeForce GTX 960 OEM— what is the largest AI model it can run?

The largest model in our catalogue that fits is DeciLM 6B at 5.7B parameters, compressed to Q3_K_M. It generates roughly 19.1 tokens per second and needs about 3.5 GB of the card's memory.

02

GeForce GTX 960 OEM— 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 40.4 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

GeForce GTX 960 OEM— how much memory does it have?

This card has 4 GB of GDDR5. Around a tenth is reserved by the inference runtime and the driver, leaving roughly 3.6 GB available for a model and its conversation.

04

GeForce GTX 960 OEM— what is its memory bandwidth?

Memory bandwidth reaches 112 GB/s across a bus of 128 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.

05

GeForce GTX 960 OEM— what type of memory does it use?

It uses GDDR5 clocked at 1.75 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

GeForce GTX 960 OEM— who makes it?

This is a product of NVIDIA, with the chip manufactured by TSMC, on a process of 28 nm.

07

GeForce GTX 960 OEM— when was it released?

It was released in November 2015.

08

GeForce GTX 960 OEM— how much cache does it have?

The L1 cache is 48 KB, and the L2 cache is 1 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.

09

GeForce GTX 960 OEM— does it support CUDA?

Yes. It reports CUDA compute capability 5.2, 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

GeForce GTX 960 OEM— what bus interface does it 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

GeForce GTX 960 OEM— is it 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

GeForce GTX 960 OEM— 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 4 GB drags the whole thing down, and none of the figures on this page assume it.

13

Would two GeForce GTX 960 OEM cards be twice as fast?

Capacity adds, throughput does not. Two of them give you 8 GB which lets you hold models neither could hold alone, but generation does not split that way. These figures describe one card.

14

GeForce GTX 960 OEM— which AI models can it run?

97 of the 679 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.

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