Calculate the TPS of the FirePro V9800 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
DeciLM 6B
5.7B · Q3_K_M · 23.0 tok/s
Fastest model
Gemma 4 E2B
57.1 tok/s · 5.1B
Which AI models can run on a FirePro V9800?
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.
105 models match
Calculating| Quantisation | Fit | ||||||
|---|---|---|---|---|---|---|---|
|
57.1
tok/s
34–91 · low confidence |
Gemma 4 E2B ≈ | 5.1B | Apr 2026 | 3.4 GB | 11k tokens ? | Q3_K_M | Tight |
|
48.6
tok/s
29–78 · low confidence |
Gemma 3 1B | 1B | Mar 2025 | 1.8 GB | 33k tokens | Q8_0 | Comfortable |
|
48.6
tok/s
29–78 · low confidence |
Gemma 3 QAT 1B | 1B | Apr 2025 | 1.8 GB | 33k tokens | Q8_0 | Comfortable |
|
48.6
tok/s
29–78 · low confidence |
HGRN 1B (WT 103) ≈ | 1B | Nov 2023 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
48.6
tok/s
29–78 · low confidence |
LLama 3..2 Typhoon 2 1B ≈ | 1B | Dec 2024 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
48.6
tok/s
29–78 · low confidence |
OLMo-1B ≈ | 1B | Feb 2024 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
48.6
tok/s
29–78 · low confidence |
Pythia-1b ≈ | 1B | Apr 2023 | 1.8 GB | 131k tokens ? | Q8_0 | Comfortable |
|
45.0
tok/s
27–72 · low confidence |
OpenELM-1.1B ≈ | 1.1B | May 2024 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
44.2
tok/s
27–71 · low confidence |
DeciCoder-1B ≈ | 1.1B | Aug 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
44.2
tok/s
27–71 · low confidence |
SantaCoder ≈ | 1.1B | Jan 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
44.2
tok/s
27–71 · low confidence |
TinyLlama-1.1B (1T token checkpoint) ≈ | 1.1B | Oct 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
44.2
tok/s
27–71 · low confidence |
TinyLlama-1.1B (3T token checkpoint) ≈ | 1.1B | Oct 2023 | 1.9 GB | 131k tokens ? | Q8_0 | Comfortable |
|
40.5
tok/s
24–65 · low confidence |
EXAONE 4.0 (1.2B) ≈ | 1.2B | Jul 2025 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
40.5
tok/s
24–65 · low confidence |
LFM2-1.2B ≈ | 1.2B | Jul 2025 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
40.5
tok/s
24–65 · low confidence |
MinerU2.5 ≈ | 1.2B | Sep 2025 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
40.5
tok/s
24–65 · low confidence |
Pleias 1.0 1.2B ≈ | 1.2B | Dec 2024 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
40.5
tok/s
24–65 · low confidence |
Pleias-RAG-1B ≈ | 1.2B | Apr 2025 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
39.5
tok/s
24–63 · low confidence |
Llama 3.2 1B | 1.2B | Sep 2024 | 2.2 GB | 54k tokens | Q8_0 | Comfortable |
|
39.0
tok/s
23–62 · low confidence |
MiniCPM-1.2B ≈ | 1.2B | Jun 2024 | 2.0 GB | 131k tokens ? | Q8_0 | Comfortable |
|
37.4
tok/s
22–60 · low confidence |
DeepSeek Coder 1.3B ≈ | 1.3B | Jan 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
37.4
tok/s
22–60 · low confidence |
DeepSeek-VL-1.3B ≈ | 1.3B | Mar 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
37.4
tok/s
22–60 · low confidence |
DigiRL ≈ | 1.3B | Jun 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
37.4
tok/s
22–60 · low confidence |
GLA Transformer 1.3B ≈ | 1.3B | Aug 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
37.4
tok/s
22–60 · low confidence |
Janus 1.3B ≈ | 1.3B | Oct 2024 | 2.1 GB | 131k tokens ? | Q8_0 | Comfortable |
|
37.4
tok/s
22–60 · low confidence |
Kosmos-2.5 ≈ | 1.3B | Aug 2024 | 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
FirePro V9800 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
- 147 GB/s
- Memory type
- GDDR5
- Memory bus width
- 256 bit
- Memory clock
- 1.15 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
- Cypress
- Architecture
- TeraScale 2
- Generation
- FirePro Terascale(Vx800)
- Foundry
- TSMC
- Process size
- 40 nm
- Transistors
- 2.2 billion
- Transistor density
- 6,400 K/mm²
- Die size
- 334 mm²
- Released
- 9 September 2010
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
- 850 MHz
- Boost clock
- 850 MHz
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
- 1,600
- Texture mapping units
- 80
- Render output units
- 32
- L1 cache
- 8 KB
- L2 cache
- 0.5 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.
- Single precision (FP32)
- 2.7 TFLOPS
- Double precision (FP64)
- 544 GFLOPS
- Pixel rate
- 27 GPixel/s
- Texture rate
- 68 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
- 1x 6-pin + 1x 8-pin
- Bus interface
- PCIe 2.0 x16
- Slot width
- Dual-slot
- Dimensions
- 267 mm
- Display outputs
- 6x mini-DisplayPort 1.1, 1x S-Video
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.
- DirectX
- 11.0
- OpenGL
- 4.4
- OpenCL
- 1.2
- Shader model
- 5.0
Listings
Where to buy a FirePro V9800
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
4 GB
Bandwidth
147 GB/s
Largest model
DeciLM 6B
FirePro V9800 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 147 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.
The figure is the bus width multiplied by a memory clock of 1.15 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 23.0 tokens per second.
The chip and how it was built
FirePro V9800 is built on the graphics processor Cypress, using the architecture TeraScale 2 from ATI, as part of the generation FirePro Terascale(Vx800).
The chip is manufactured by TSMC, on a process of 40 nm, with a die measuring 334 mm², holding 2.2 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 September 2010, roughly 16.01284980322 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
544 GFLOPS
Double-precision throughput reaches 544 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 850 MHz to a boost of 850 MHz. 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
FirePro V9800 has an L1 cache of 8 KB, backed by an L2 cache of 0.5 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 1,600 shading units, 80 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
250 W
FirePro V9800 is rated at 250 W, and the suggested system power supply is 600 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 dual-slot, measuring 267 mm long, and needs 1x 6-pin + 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 2.0 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 FirePro V9800
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 FirePro V9800
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 FirePro V9800
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 105 models this card can run. Search by name, or by size — typing 27b matches on the parameter count even when the name never states it.
-
02
Decide how long your conversations run
Drag the slider to the conversation length you plan to work at. The cache grows with the conversation, and against a card holding 4 GB so the setting is worth getting right.
-
03
Set a minimum quality if you need one
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
Look at the range, not just the number
Speeds come with error bars for a reason. The best case here is 57.1 tok/s on Gemma 4 E2B. The same card and model vary by thirty to fifty per cent between inference runtimes.
-
05
Check the memory column before committing
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 4 GB.
-
06
Cross-check against other hardware
Following a model through to its own page lists all the hardware that can run it, so you can see how it compares against FirePro V9800.
Answers
FirePro V9800 — common questions
FirePro V9800— which AI models can it run?
105 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.
FirePro V9800— 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 23.0 tokens per second and needs about 3.5 GB of the card's memory.
FirePro V9800— how many tokens per second does it produce?
It depends on the model. The fastest model we track here is Gemma 4 E2B at about 57.1 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.
FirePro V9800— 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.
FirePro V9800— what is its memory bandwidth?
Memory bandwidth reaches 147 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.
FirePro V9800— what type of memory does it use?
It uses GDDR5 clocked at 1.15 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.
FirePro V9800— who makes it?
This is a product of ATI, with the chip manufactured by TSMC, on a process of 40 nm.
FirePro V9800— when was it released?
It was released in September 2010.
FirePro V9800— how much power does it use?
Rated board power is 250 W, and the suggested system power supply is 600 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.
FirePro V9800— how much cache does it have?
The L1 cache is 8 KB, and the L2 cache is 0.5 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.
FirePro V9800— does it support CUDA?
No. CUDA is NVIDIA-only, and this is a card from ATI. It runs language models through ROCm, Vulkan or Metal depending on the software, which are less mature than the CUDA path — our estimates apply a penalty for that.
FirePro V9800— what bus interface does it use?
It uses PCIe 2.0 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.
FirePro V9800— 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 105 of the models we track. Whether that is enough depends entirely on which model you want — the table above answers that directly.
FirePro V9800— 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 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 FirePro V9800 cards be twice as fast?
Capacity adds, throughput does not. Two of them give you 8 GB to work with rather than twice the tokens per second — every figure here is for a single card.
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.