Calculate the TPS of the CMP 30HX 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
Qwen-VL
9.6B · Q3_K_M · 40.0 tok/s
Fastest model
Gemma 3 QAT 1B
142 tok/s · 1B
Which AI models can run on a CMP 30HX?
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.
266 models match
Calculating| Quantisation | Fit | ||||||
|---|---|---|---|---|---|---|---|
|
142
tok/s
121–171 |
Gemma 3 1B | 1B | Mar 2025 | 1.8 GB | 33k tokens | Q8_0 | Comfortable |
|
142
tok/s
121–171 |
Gemma 3 QAT 1B | 1B | Apr 2025 | 1.8 GB | 33k tokens | Q8_0 | Comfortable |
|
142
tok/s
85–228 · low confidence |
HGRN 1B (WT 103) ≈ | 1B | Nov 2023 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
142
tok/s
85–228 · low confidence |
LLama 3..2 Typhoon 2 1B ≈ | 1B | Dec 2024 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
142
tok/s
85–228 · low confidence |
OLMo-1B ≈ | 1B | Feb 2024 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
142
tok/s
85–228 · low confidence |
Pythia-1b ≈ | 1B | Apr 2023 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
132
tok/s
79–211 · low confidence |
OpenELM-1.1B ≈ | 1.1B | May 2024 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
129
tok/s
78–207 · low confidence |
DeciCoder-1B ≈ | 1.1B | Aug 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
129
tok/s
78–207 · low confidence |
SantaCoder ≈ | 1.1B | Jan 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
129
tok/s
78–207 · low confidence |
TinyLlama-1.1B (1T token checkpoint) ≈ | 1.1B | Oct 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
129
tok/s
78–207 · low confidence |
TinyLlama-1.1B (3T token checkpoint) ≈ | 1.1B | Oct 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
119
tok/s
71–190 · low confidence |
EXAONE 4.0 (1.2B) ≈ | 1.2B | Jul 2025 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
119
tok/s
71–190 · low confidence |
MinerU2.5 ≈ | 1.2B | Sep 2025 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
119
tok/s
71–190 · low confidence |
Pleias 1.0 1.2B ≈ | 1.2B | Dec 2024 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
119
tok/s
71–190 · low confidence |
Pleias-RAG-1B ≈ | 1.2B | Apr 2025 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
116
tok/s
98–139 |
Llama 3.2 1B | 1.2B | Sep 2024 | 2.2 GB | 113k tokens | Q8_0 | Comfortable |
|
114
tok/s
68–183 · low confidence |
MiniCPM-1.2B ≈ | 1.2B | Jun 2024 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
109
tok/s
66–175 · low confidence |
DeepSeek Coder 1.3B ≈ | 1.3B | Jan 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
109
tok/s
66–175 · low confidence |
DeepSeek-VL-1.3B ≈ | 1.3B | Mar 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
109
tok/s
66–175 · low confidence |
DigiRL ≈ | 1.3B | Jun 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
109
tok/s
66–175 · low confidence |
GLA Transformer 1.3B ≈ | 1.3B | Aug 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
109
tok/s
66–175 · low confidence |
Janus 1.3B ≈ | 1.3B | Oct 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
109
tok/s
66–175 · low confidence |
Kosmos-2.5 ≈ | 1.3B | Aug 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
109
tok/s
66–175 · low confidence |
Otter ≈ | 1.3B | May 2023 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
109
tok/s
66–175 · 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
CMP 30HX 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
- 6 GB
- Memory bandwidth
- 336 GB/s
- Memory type
- GDDR6
- Memory bus width
- 192 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
- TU116
- Architecture
- Turing
- Generation
- Mining GPUs
- Foundry
- TSMC
- Process size
- 12 nm
- Transistors
- 6.6 billion
- Transistor density
- 23,200 K/mm²
- Die size
- 284 mm²
- Package
- BGA-2228
- Released
- 25 February 2021
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
- 1.53 GHz
- Boost clock
- 1.79 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
- 1,408
- Texture mapping units
- 88
- Render output units
- 48
- Streaming multiprocessors
- 22
- L1 cache
- 64 KB
- L2 cache
- 1.5 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)
- 10.1 TFLOPS
- Single precision (FP32)
- 5 TFLOPS
- Double precision (FP64)
- 157.1 GFLOPS
- Pixel rate
- 86 GPixel/s
- Texture rate
- 157 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)
- 125 W
- Suggested power supply
- 300 W
- Power connectors
- 1x 8-pin
- Bus interface
- PCIe 1.0 x4
- Slot width
- Dual-slot
- Dimensions
- 229 mm × 35 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
- 7.5
- DirectX
- 12.1
- OpenGL
- 4.6
- Vulkan
- 1.4
- OpenCL
- 3.0
- Shader model
- 6.8
Listings
Where to buy a CMP 30HX
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
Memory: the specification that decides everything
Memory
6 GB
Bandwidth
336 GB/s
Largest model
Qwen-VL
At 6 GB of GDDR6 the CMP 30HX is limited to the smaller end of the catalogue. About 5.4 GB is actually available to a runtime, and a model has to fit entirely inside it before generating anything at all.
The memory bus moves 336 GB/s across a 192-bit bus. That is the number that governs generation speed — arithmetic per byte read is small enough that the bus, not the cores, is what everything waits on.
Bandwidth is clock times bus width, and this card clocks its memory at 1.75 GHz. Both halves matter, and neither is visible in a gaming benchmark.
Put together, the largest model that fits is Qwen-VL at 9.6B, running Q3_K_M and producing around 40.0 tokens per second.
The chip and how it was built
The CMP 30HX is built on the TU116 graphics processor, using NVIDIA's Turing architecture, as part of the Mining GPUs generation.
The chip is manufactured by TSMC, on a 12 nm process, with a die measuring 284 mm², holding 6.6 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 February 2021, roughly 5 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
10.1 TFLOPS
FP64
157.1 GFLOPS
On paper the CMP 30HX reaches 10.1 TFLOPS at half precision and 5 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 157.1 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 1.53 GHz at base to 1.79 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 CMP 30HX has 64 KB of L1 cache, backed by 1.5 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 1,408 shading units, 88 texture mapping units, and 48 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
125 W
The CMP 30HX is rated at 125 W, with a 300 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, measuring 229 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 1.0 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 CMP 30HX
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 CMP 30HX
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 CMP 30HX
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
Search for the model you want
Every one of the 266 models this CMP 30HX runs is in the table above. Search narrows it by name or by size.
-
02
Match the context to your work
Set the context to your real working length. Short questions cost almost nothing; a long document can consume a large share of the card's 6 GB.
-
03
Pin the comparison to one quality level
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.
-
04
Read the speed and the range
Each speed is an estimate for a single conversation, with a range beneath it — 142 tok/s on Gemma 3 QAT 1B at the top end here. The same card and model vary by thirty to fifty per cent between inference runtimes.
-
05
Check the headroom before you decide
Compare what each model needs with the 6 GB this card provides. Tight means it works today; comfortable means it still works when the conversation grows.
-
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, and how the CMP 30HX compares.
Answers
CMP 30HX — common questions
Who makes the CMP 30HX?
The CMP 30HX is a NVIDIA product, with the chip manufactured by TSMC, on a 12 nm process.
When was the CMP 30HX released?
The CMP 30HX was released in February 2021.
How much power does a CMP 30HX use?
The CMP 30HX has a rated board power of 125 W, and a 300 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 CMP 30HX have?
The CMP 30HX has 64 KB of L1 cache, and 1.5 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 CMP 30HX?
The CMP 30HX is rated at 10.1 TFLOPS at half precision and 5 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.
Does the CMP 30HX support CUDA?
Yes. The CMP 30HX reports CUDA compute capability 7.5. 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.
What bus interface does the CMP 30HX use?
It uses PCIe 1.0 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.
Is the CMP 30HX 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 266 of the models we track. Whether that is enough depends entirely on which model you want — the table above answers that directly.
Can a CMP 30HX run a model that does not fit in its memory?
Only partly. Layers beyond the 6 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 CMP 30HX cards be twice as fast?
No. A second CMP 30HX doubles the memory to 12 GB, which lets you hold models neither could hold alone, but generation does not split that way. These figures describe one card.
What AI models can a CMP 30HX run?
266 of the 679 open-weight language models we track fit on a CMP 30HX 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 CMP 30HX can run?
The largest model in our catalogue that fits on a CMP 30HX is Qwen-VL at 9.6B parameters, compressed to Q3_K_M. It generates roughly 40.0 tokens per second and needs about 5.4 GB of the card's memory.
How many tokens per second does a CMP 30HX produce?
It depends on the model. On a CMP 30HX the fastest model we track is Gemma 3 QAT 1B at about 142 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 CMP 30HX run a 7B model?
Yes. For example a CMP 30HX runs MetaMath 7B (Mistral finetune) at IQ4_XS, using about 5.1 GB of memory and generating around 49.9 tokens per second.
How much memory does a CMP 30HX have?
A CMP 30HX has 6 GB of GDDR6 memory. Around a tenth of that is reserved by the inference runtime and the driver, leaving roughly 5.4 GB available for a model and its conversation.
What is the memory bandwidth of a CMP 30HX?
The CMP 30HX has 336 GB/s of memory bandwidth, across a 192-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 CMP 30HX use?
It uses GDDR6 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.
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.