Calculate the TPS of the Jetson AGX Orin 64 GB 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
721 models in our catalogue altogether
Largest model it holds
Qwen3.5-122B-A10B
122B · Q3_K_M · 10.7 tok/s
Fastest model
Gemma 3 QAT 1B
86.7 tok/s · 1B
Which AI models can run on a Jetson AGX Orin 64 GB?
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.
628 models match
Calculating| Quantisation | Fit | ||||||
|---|---|---|---|---|---|---|---|
|
86.7
tok/s
74–104 |
Gemma 3 1B | 1B | Mar 2025 | 1.8 GB | 33k tokens | Q8_0 | Comfortable |
|
86.7
tok/s
74–104 |
Gemma 3 QAT 1B | 1B | Apr 2025 | 1.8 GB | 33k tokens | Q8_0 | Comfortable |
|
86.7
tok/s
52–139 · low confidence |
HGRN 1B (WT 103) ≈ | 1B | Nov 2023 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
86.7
tok/s
52–139 · low confidence |
LLama 3..2 Typhoon 2 1B ≈ | 1B | Dec 2024 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
86.7
tok/s
52–139 · low confidence |
OLMo-1B ≈ | 1B | Feb 2024 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
86.7
tok/s
52–139 · low confidence |
Pythia-1b ≈ | 1B | Apr 2023 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
80.3
tok/s
48–129 · low confidence |
OpenELM-1.1B ≈ | 1.1B | May 2024 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
78.9
tok/s
47–126 · low confidence |
DeciCoder-1B ≈ | 1.1B | Aug 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
78.9
tok/s
47–126 · low confidence |
SantaCoder ≈ | 1.1B | Jan 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
78.9
tok/s
47–126 · low confidence |
TinyLlama-1.1B (1T token checkpoint) ≈ | 1.1B | Oct 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
78.9
tok/s
47–126 · low confidence |
TinyLlama-1.1B (3T token checkpoint) ≈ | 1.1B | Oct 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
72.3
tok/s
43–116 · low confidence |
EXAONE 4.0 (1.2B) ≈ | 1.2B | Jul 2025 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
72.3
tok/s
43–116 · low confidence |
LFM2-1.2B ≈ | 1.2B | Jul 2025 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
72.3
tok/s
43–116 · low confidence |
MinerU2.5 ≈ | 1.2B | Sep 2025 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
72.3
tok/s
43–116 · low confidence |
Pleias 1.0 1.2B ≈ | 1.2B | Dec 2024 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
72.3
tok/s
43–116 · low confidence |
Pleias-RAG-1B ≈ | 1.2B | Apr 2025 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
70.5
tok/s
60–85 |
Llama 3.2 1B | 1.2B | Sep 2024 | 2.2 GB | 131k tokens | Q8_0 | Comfortable |
|
69.5
tok/s
42–111 · low confidence |
MiniCPM-1.2B ≈ | 1.2B | Jun 2024 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
66.7
tok/s
40–107 · low confidence |
DeepSeek Coder 1.3B ≈ | 1.3B | Jan 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
66.7
tok/s
40–107 · low confidence |
DeepSeek-VL-1.3B ≈ | 1.3B | Mar 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
66.7
tok/s
40–107 · low confidence |
DigiRL ≈ | 1.3B | Jun 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
66.7
tok/s
40–107 · low confidence |
GLA Transformer 1.3B ≈ | 1.3B | Aug 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
66.7
tok/s
40–107 · low confidence |
Janus 1.3B ≈ | 1.3B | Oct 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
66.7
tok/s
40–107 · low confidence |
Kosmos-2.5 ≈ | 1.3B | Aug 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
66.7
tok/s
40–107 · 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
Jetson AGX Orin 64 GB 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
- 64 GB
- Memory bandwidth
- 205 GB/s
- Memory type
- LPDDR5
- Memory bus width
- 256 bit
- Memory clock
- 800 MHz
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
- GA10B
- Architecture
- Ampere
- Generation
- Tegra(Ampere)
- Foundry
- Samsung
- Process size
- 8 nm
- Die size
- 200 mm²
- Released
- 1 March 2023
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
- 1.3 GHz
- Boost clock
- 1.3 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
- 2,048
- Texture mapping units
- 64
- Render output units
- 32
- Streaming multiprocessors
- 16
- Tensor cores
- 64
- 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)
- 10.7 TFLOPS
- Single precision (FP32)
- 5.3 TFLOPS
- Pixel rate
- 42 GPixel/s
- Texture rate
- 83 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)
- 60 W
- Bus interface
- PCIe 4.0 x4
- Slot width
- IGP
- Dimensions
- 100 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
- 8.7
- DirectX
- 12.2
- OpenGL
- 4.6
- Vulkan
- 1.4
- OpenCL
- 3.0
- Shader model
- 6.8
Listings
Where to buy a Jetson AGX Orin 64 GB
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
64 GB
Bandwidth
205 GB/s
Largest model
Qwen3.5-122B-A10B
Jetson AGX Orin 64 GB carries 64 GB of LPDDR5. That covers the mid-sized models most people actually run. Once the runtime and driver reserve their working space, roughly this much is left: 57.6 GB.
Memory bandwidth reaches 205 GB/s across a bus of 256 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.
Bandwidth is clock times bus width, and this card clocks its memory at 800 MHz. Both halves matter, and neither is visible in a gaming benchmark.
In practice that combination tops out at Qwen3.5-122B-A10B, 122B, compressed to Q3_K_M and generating around 10.7 tokens per second.
The chip and how it was built
Jetson AGX Orin 64 GB is built on the graphics processor GA10B, using the architecture Ampere from NVIDIA, as part of the generation Tegra(Ampere).
The chip is manufactured by Samsung, on a process of 8 nm, with a die measuring 200 mm². 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 March 2023, roughly 3.5378175655534 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
10.7 TFLOPS
Tensor cores
64
On paper Jetson AGX Orin 64 GB reaches 10.7 TFLOPS at half precision, and 5.3 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.
The card carries 64 tensor cores across 16 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 1.3 GHz to a boost of 1.3 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
Jetson AGX Orin 64 GB has an L1 cache of 128 KB, backed by an L2 cache of 6 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 2,048 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
Power draw
60 W
Jetson AGX Orin 64 GB is rated at 60 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, measuring 100 mm long. 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 x4. 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 Jetson AGX Orin 64 GB
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 Jetson AGX Orin 64 GB
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 Jetson AGX Orin 64 GB
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
Find the model in the table
The table lists 628 models this card runs. Search narrows the list by name or by size.
-
02
Match the context to your work
Drag the slider to the conversation length you plan to work at. The cache grows with the conversation, and against a card holding 64 GB 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
Speeds come with error bars for a reason. The best case here is 86.7 tok/s on Gemma 3 QAT 1B. The same card and model vary by thirty to fifty per cent between inference runtimes.
-
05
Check the headroom before you decide
A tight fit runs but leaves no room to raise the context later; comfortable has headroom. The memory column shows what each model needs against an available 64 GB.
-
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, alongside Jetson AGX Orin 64 GB.
Answers
Jetson AGX Orin 64 GB — common questions
Jetson AGX Orin 64 GB— is it good for running local AI models?
Its memory is large enough for models most desktop hardware cannot touch though its bandwidth means generation will feel slow on larger models. In total it runs 628 of the models we track. Whether that is enough depends entirely on which model you want — the table above answers that directly.
Jetson AGX Orin 64 GB— can it run a model that does not fit in its memory?
Offloading past the card's 64 GB is possible and usually a false economy: the system-memory portion is slow enough to dominate the result.
Would two Jetson AGX Orin 64 GB cards be twice as fast?
Pairing them buys headroom rather than pace: 128 GB to work with rather than twice the tokens per second — every figure here is for a single card.
Jetson AGX Orin 64 GB— which AI models can it run?
628 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.
Jetson AGX Orin 64 GB— what is the largest AI model it can run?
The largest model in our catalogue that fits is Qwen3.5-122B-A10B at 122B parameters, compressed to Q3_K_M. It generates roughly 10.7 tokens per second and needs about 53.1 GB of the card's memory.
Jetson AGX Orin 64 GB— 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 86.7 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.
Jetson AGX Orin 64 GB— 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 19.3 tokens per second.
Jetson AGX Orin 64 GB— 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 30.1 tokens per second.
Jetson AGX Orin 64 GB— 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 17.2 tokens per second.
Jetson AGX Orin 64 GB— can it run 70B models?
Yes. For example it runs Qwen3-Coder-Next at Q5_K_M, using about 52.7 GB of memory and generating around 10.8 tokens per second.
Jetson AGX Orin 64 GB— how much memory does it have?
This card has 64 GB of LPDDR5. Around a tenth is reserved by the inference runtime and the driver, leaving roughly 57.6 GB available for a model and its conversation.
Jetson AGX Orin 64 GB— what is its memory bandwidth?
Memory bandwidth reaches 205 GB/s across a bus of 256 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.
Jetson AGX Orin 64 GB— what type of memory does it use?
It uses LPDDR5 clocked at 800 MHz. 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.
Jetson AGX Orin 64 GB— who makes it?
This is a product of NVIDIA, with the chip manufactured by Samsung, on a process of 8 nm.
Jetson AGX Orin 64 GB— when was it released?
It was released in March 2023.
Jetson AGX Orin 64 GB— how much power does it use?
Rated board power is 60 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.
Jetson AGX Orin 64 GB— how much cache does it have?
The L1 cache is 128 KB, and the L2 cache is 6 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.
Jetson AGX Orin 64 GB— what are its TFLOPS?
It is rated at 10.7 TFLOPS at half precision and 5.3 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.
Jetson AGX Orin 64 GB— how many tensor cores does it have?
It has 64 tensor cores across 16 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.
Jetson AGX Orin 64 GB— does it support CUDA?
Yes. It reports CUDA compute capability 8.7. 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.
Jetson AGX Orin 64 GB— what bus interface does it use?
It uses PCIe 4.0 x4. 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.
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.