Calculate the TPS of the Tesla PG503-216 on local AI models
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
679 models in our catalogue altogether
Largest model it holds
Phi-3.5-MoE
60.8B · Q3_K_M · 118 tok/s
Fastest model
Gemma 3 QAT 1B
479 tok/s · 1B
Which AI models can run on a Tesla PG503-216?
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.
513 models match
Calculating| Quantisation | Fit | ||||||
|---|---|---|---|---|---|---|---|
|
479
tok/s
407–574 |
Gemma 3 1B | 1B | Mar 2025 | 1.8 GB | 33k tokens | Q8_0 | Comfortable |
|
479
tok/s
407–574 |
Gemma 3 QAT 1B | 1B | Apr 2025 | 1.8 GB | 33k tokens | Q8_0 | Comfortable |
|
479
tok/s
287–766 · low confidence |
HGRN 1B (WT 103) ≈ | 1B | Nov 2023 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
479
tok/s
287–766 · low confidence |
LLama 3..2 Typhoon 2 1B ≈ | 1B | Dec 2024 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
479
tok/s
287–766 · low confidence |
OLMo-1B ≈ | 1B | Feb 2024 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
479
tok/s
287–766 · low confidence |
Pythia-1b ≈ | 1B | Apr 2023 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
443
tok/s
266–709 · low confidence |
OpenELM-1.1B ≈ | 1.1B | May 2024 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
435
tok/s
261–696 · low confidence |
DeciCoder-1B ≈ | 1.1B | Aug 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
435
tok/s
261–696 · low confidence |
SantaCoder ≈ | 1.1B | Jan 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
435
tok/s
261–696 · low confidence |
TinyLlama-1.1B (1T token checkpoint) ≈ | 1.1B | Oct 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
435
tok/s
261–696 · low confidence |
TinyLlama-1.1B (3T token checkpoint) ≈ | 1.1B | Oct 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
399
tok/s
239–638 · low confidence |
EXAONE 4.0 (1.2B) ≈ | 1.2B | Jul 2025 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
399
tok/s
239–638 · low confidence |
MinerU2.5 ≈ | 1.2B | Sep 2025 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
399
tok/s
239–638 · low confidence |
Pleias 1.0 1.2B ≈ | 1.2B | Dec 2024 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
399
tok/s
239–638 · low confidence |
Pleias-RAG-1B ≈ | 1.2B | Apr 2025 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
389
tok/s
331–467 |
Llama 3.2 1B | 1.2B | Sep 2024 | 2.2 GB | 131k tokens | Q8_0 | Comfortable |
|
384
tok/s
230–614 · low confidence |
MiniCPM-1.2B ≈ | 1.2B | Jun 2024 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
368
tok/s
221–589 · low confidence |
DeepSeek Coder 1.3B ≈ | 1.3B | Jan 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
368
tok/s
221–589 · low confidence |
DeepSeek-VL-1.3B ≈ | 1.3B | Mar 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
368
tok/s
221–589 · low confidence |
DigiRL ≈ | 1.3B | Jun 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
368
tok/s
221–589 · low confidence |
GLA Transformer 1.3B ≈ | 1.3B | Aug 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
368
tok/s
221–589 · low confidence |
Janus 1.3B ≈ | 1.3B | Oct 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
368
tok/s
221–589 · low confidence |
Kosmos-2.5 ≈ | 1.3B | Aug 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
368
tok/s
221–589 · low confidence |
Otter ≈ | 1.3B | May 2023 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
368
tok/s
221–589 · 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
Tesla PG503-216 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
- 32 GB
- Memory bandwidth
- 1,130 GB/s
- Memory type
- HBM2
- Memory bus width
- 4,096 bit
- Memory clock
- 1.11 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
- GV100
- Architecture
- Volta
- Generation
- Tesla Volta(Vxx)
- Foundry
- TSMC
- Process size
- 12 nm
- Transistors
- 21.1 billion
- Transistor density
- 25,900 K/mm²
- Die size
- 815 mm²
- Released
- 26 November 2019
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
- 1.31 GHz
- Boost clock
- 1.53 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
- 4,608
- Texture mapping units
- 288
- Render output units
- 128
- Streaming multiprocessors
- 80
- Tensor cores
- 640
- L1 cache
- 128 KB
- L2 cache
- 6 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)
- 28.2 TFLOPS
- Single precision (FP32)
- 14.1 TFLOPS
- Double precision (FP64)
- 7.1 TFLOPS
- Pixel rate
- 196 GPixel/s
- Texture rate
- 441 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)
- 250 W
- Suggested power supply
- 600 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
- 7.0
- DirectX
- 12.1
- OpenGL
- 4.6
- Vulkan
- 1.4
- OpenCL
- 3.0
- Shader model
- 6.8
Listings
Where to buy a Tesla PG503-216
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
32 GB
Bandwidth
1,130 GB/s
Largest model
Phi-3.5-MoE
The Tesla PG503-216 carries 32 GB of HBM2, which covers the mid-sized models most people actually run — about 28.8 GB of it after the runtime and driver reserve their working space.
Bandwidth is 1,130 GB/s across a 4,096-bit bus. 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 1.11 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.
The biggest thing it holds is Phi-3.5-MoE (60.8B) at Q3_K_M compression, for about 118 tokens per second.
The chip and how it was built
The Tesla PG503-216 is built on the GV100 graphics processor, using NVIDIA's Volta architecture, as part of the Tesla Volta(Vxx) generation.
The chip is manufactured by TSMC, on a 12 nm process, with a die measuring 815 mm², holding 21.1 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 2019, roughly 6 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
28.2 TFLOPS
FP64
7.1 TFLOPS
Tensor cores
640
On paper the Tesla PG503-216 reaches 28.2 TFLOPS at half precision and 14.1 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 is 7.1 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 640 tensor cores across 80 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 1.31 GHz at base to 1.53 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 Tesla PG503-216 has 128 KB of L1 cache, backed by 6 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 4,608 shading units, 288 texture mapping units, and 128 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
250 W
The Tesla PG503-216 is rated at 250 W, with a 600 W power supply suggested for the whole system. 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 a 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 Tesla PG503-216
The biggest open-weight models that fit on this card, newest first. Each is shown at the best compression the card can hold.
The fastest AI models on a Tesla PG503-216
Where this card produces tokens quickest. Smaller models dominate here, because generating each token means reading the whole model out of memory once.
Step by step
How to work out the tokens per second of a Tesla PG503-216
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.
-
01
Start with the model, not the specification
The table lists 513 models this Tesla PG503-216 can run. Search by name, or by size — typing 27b matches on the parameter count even when the name never states it.
-
02
Match the context to your work
Longer conversations cost memory on top of the weights. With 32 GB to work in, that is frequently the difference between a model fitting and not.
-
03
Pin the comparison to one quality level
Compression is what lets bigger models fit. The quality control drops any model that needs more of it than you are willing to give.
-
04
Read the speed and the range
The figures are calculated, not measured. 479 tok/s on Gemma 3 QAT 1B is the fastest result on this card, and like every row it carries a range that reflects how much the runtime matters.
-
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 32 GB before settling on one.
-
06
Cross-check against other hardware
Every model name in the table links to its own page, which runs the same calculation across every card we hold. That is where you see whether the Tesla PG503-216 is the right buy for it or merely a card that fits.
Answers
Tesla PG503-216 — common questions
What bus interface does the Tesla PG503-216 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.
Is the Tesla PG503-216 good for running local AI models?
Its memory comfortably covers the mid-sized models most people run locally and its bandwidth is high enough to generate text faster than most people read. In total it runs 513 of the models we track. Whether that is enough depends entirely on which model you want — the table above answers that directly.
Can a Tesla PG503-216 run a model that does not fit in its memory?
Only partly. Layers beyond the 32 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.
Would two Tesla PG503-216 cards be twice as fast?
Capacity adds, throughput does not. Two of them give you 64 GB to work with rather than twice the tokens per second — every figure here is for a single Tesla PG503-216.
What AI models can a Tesla PG503-216 run?
513 of the 679 open-weight language models we track fit on a Tesla PG503-216 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.
What is the largest AI model a Tesla PG503-216 can run?
The largest model in our catalogue that fits on a Tesla PG503-216 is Phi-3.5-MoE at 60.8B parameters, compressed to Q3_K_M. It generates roughly 118 tokens per second and needs about 27.2 GB of the card's memory.
How many tokens per second does a Tesla PG503-216 produce?
It depends on the model. On a Tesla PG503-216 the fastest model we track is Gemma 3 QAT 1B at about 479 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.
Can a Tesla PG503-216 run a 7B model?
Yes. For example a Tesla PG503-216 runs Multi-Token Prediction 7B at Q8_0, using about 7.9 GB of memory and generating around 71.4 tokens per second.
Can a Tesla PG503-216 run a 13B model?
Yes. For example a Tesla PG503-216 runs DeepSeekMoE-16B at Q8_0, using about 17.5 GB of memory and generating around 166 tokens per second.
Can a Tesla PG503-216 run a 30B model?
Yes. For example a Tesla PG503-216 runs ERNIE-4.5-VL-28B-A3B at Q6_K, using about 22.7 GB of memory and generating around 138 tokens per second.
How much memory does a Tesla PG503-216 have?
A Tesla PG503-216 has 32 GB of HBM2 memory. Around a tenth of that is reserved by the inference runtime and the driver, leaving roughly 28.8 GB available for a model and its conversation.
What is the memory bandwidth of a Tesla PG503-216?
The Tesla PG503-216 has 1,130 GB/s of memory bandwidth, across a 4,096-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.
What type of memory does a Tesla PG503-216 use?
It uses HBM2 clocked at 1.11 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.
Who makes the Tesla PG503-216?
The Tesla PG503-216 is a NVIDIA product, with the chip manufactured by TSMC, on a 12 nm process.
When was the Tesla PG503-216 released?
The Tesla PG503-216 was released in November 2019.
How much power does a Tesla PG503-216 use?
The Tesla PG503-216 has a rated board power of 250 W, and a 600 W system power supply is suggested. Generating text draws hard in bursts and idles between requests, so average consumption over a working session is normally well below the rated figure.
How much cache does a Tesla PG503-216 have?
The Tesla PG503-216 has 128 KB of L1 cache, and 6 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.
What are the TFLOPS of a Tesla PG503-216?
The Tesla PG503-216 is rated at 28.2 TFLOPS at half precision and 14.1 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.
How many tensor cores does a Tesla PG503-216 have?
The Tesla PG503-216 has 640 tensor cores across 80 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.
Does the Tesla PG503-216 support CUDA?
Yes. The Tesla PG503-216 reports CUDA compute capability 7.0. 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.
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.