NVIDIA A100 80GBコスパ重視80 GB VRAM (+32)2039 GB/s (+1367)
C
Raises estimated decode speed by about 269%.127.6 tok/sデコード
Raises estimated decode speed by about 269%.
Adds memory headroom for longer context windows and future model growth.
〜$15,000 MSRP
Can Codestral 22B v0.1 IMat run on Quadro RTX 8000 48GB?
YES — Runs Great
C48Usable
Estimated from fit model
Codestral 22B v0.1 IMat needs ~22.0 GB VRAM. Quadro RTX 8000 48GB has 48.0 GB. With Q4_K_M quantization, expect ~35 tok/s.
Runtime: OllamaCapacity: RoomyBandwidth: MediumStack: BasicBottleneck: Balanced
Operating mode
Choose the run profile you care about
Interactive favors responsiveness, while light API and scale-out lean harder on serving readiness. The fit stays the same, but the recommendation lens changes.
Current mode
Balanced
Balanced for general local use. Keeps the ranking neutral across personal and serving workflows.
Select quantization to explore
Q4_K_M (Medium quality) — 22.0 GB, 34.6 tok/s, Runs well
22.0 GB required48.0 GB available
46% VRAM used
Fit status
Runs well
Decode
34.6 tok/s
TTFT
5603 ms
Safe context
177K
Memory
22.0 GB / 48.0 GB
Memory breakdown
Weights13.4 GB
KV Cache2.6 GB
Runtime1.2 GB
Headroom4.8 GB
See how fast it feels
See how fast it feelsCodestral 22B v0.1 IMat on Quadro RTX 8000 48GB
1st promptCold start — includes initialization
>Write a Python function to sort a list
Here's a Python function that sorts a list using quicksort:
def quicksort(arr):
if len(arr) <= 1:
return arr
pivot = arr[len(arr) // 2]
left = [x for x in arr if x < pivot]
middle = [x for x in arr if x == pivot]
right = [x for x in arr if x > pivot]
return quicksort(left) + middle + quicksort(right)
# Example usage
numbers = [3, 6, 8, 10, 1, 2, 1]
print(quicksort(numbers)) # [1, 1, 2, 3, 6, 8, 10]
The algorithm picks a pivot, partitions the list into smaller, equal, and larger elements, then recursively sorts each partition and concatenates the results.
2nd promptWarm — model ready, no init overhead
>Now add type hints and a docstring
def quicksort(arr: list[int]) -> list[int]:
"""Sort a list using the quicksort algorithm.
Args:
arr: List of integers to sort.
Returns:
A new sorted list.
"""
if len(arr) <= 1:
return arr
pivot = arr[len(arr) // 2]
left = [x for x in arr if x < pivot]
middle = [x for x in arr if x == pivot]
right = [x for x in arr if x > pivot]
return quicksort(left) + middle + quicksort(right)
Added type hints for the input and return type, plus a Google-style docstring.
Estimated: 34.6 tok/s decode · 5.6s TTFT (warm) · 86 tok/s prefill
What limits this setup
This setup is broadly balanced for this model.
Older PCIe generation
PCIe 3.0 is workable, but it compounds the penalty when you offload heavily or try to scale across multiple cards.
Best improvement path
Performance by workload
| Workload | Grade | Fit | Decode | TTFT | Context |
|---|---|---|---|---|---|
| Chat | C | Runs well | 34.6 tok/s | 3056 ms | 177K |
| Coding | C | Runs well | 34.6 tok/s | 5603 ms | 177K |
| Agentic Coding | C | Runs well | 34.6 tok/s | 8150 ms | 177K |
| Reasoning | C | Runs well | 34.6 tok/s | 6622 ms | 177K |
| RAG | C | Runs well | 34.6 tok/s | 10188 ms | 177K |
Inference speed
Codestral 22B v0.1 IMat inference speed — tokens per second by GPU & Mac
Estimated decode speed (tokens/sec) for Codestral 22B v0.1 IMat at Q4_K_M across popular GPUs and Apple Silicon, using the fastest local runtime per device. Fastest is RTX 5090 32GB at ~90 tok/s. Speed is memory-bandwidth bound, so cards that fit the whole model in VRAM run far faster than ones that offload to system RAM.
| GPU / Mac | Memory | Quant | Speed (tok/s) | Fits? |
|---|---|---|---|---|
| RTX 5090 32GB | 32 GB | Q4_K_M | 89.5 | Fits |
| RTX 4090 24GB | 24 GB | Q4_K_M | 57.1 | Fits |
RX 7900 XTX 24GB | 24 GB | Q4_K_M | 51.5 | Fits |
| RTX 3090 24GB | 24 GB | Q4_K_M | 48.8 | Fits |
Mac Studio M3 Ultra 256GB | 256 GB | Q4_K_M | 41.5 | Fits |
MacBook Pro M4 Max 128GB | 128 GB | Q4_K_M | 34.8 | Fits |
MacBook Pro M4 Max 64GB | 64 GB | Q4_K_M | 34.8 | Fits |
Mac Studio M2 Ultra 128GB | 128 GB | Q4_K_M | 34.6 | Fits |
Mac Studio M1 Ultra 128GB | 128 GB | Q4_K_M | 32.8 | Fits |
| RTX 4080 Super 16GB | 16 GB | Q4_K_M | 24.3 | Heavy offload |
MacBook Pro M4 Pro 48GB | 48 GB | Q4_K_M | 21.9 | Fits |
MacBook Pro M3 Max 64GB | 64 GB | Q4_K_M | 17.9 | Fits |
MacBook Pro M1 Max 64GB | 64 GB | Q4_K_M | 16.4 | Fits |
| RTX 4070 12GB | 12 GB | Q4_K_M | 8.6 | Too big |
| RTX 3060 12GB | 12 GB | Q4_K_M | 5.4 | Too big |
| RTX 4060 8GB | 8 GB | Q4_K_M | 2.2 | Too big |
Estimates for single-stream decoding at Q4_K_M; real tokens/sec varies with prompt length, context, batch size, and runtime build. Prompt processing (prefill) is faster than the decode figures shown here.
Quantization options
How Codestral 22B v0.1 IMat (22B params) fits at each quantization level on Quadro RTX 8000 48GB (48.0 GB usable).
| Quant | Bits | VRAM | Quality | Fit |
|---|---|---|---|---|
Q2_K | 2 | 8.6 GB | Low | C42 |
Q3_K_S | 3 | 10.8 GB | Low | C43 |
NVFP4 | 4 | 12.3 GB | Medium | C43 |
Q4_K_M | 4 | 13.4 GB | Medium | C44 |
Q5_K_M | 5 | 15.8 GB | High | C44 |
Q6_K | 6 | 18.0 GB | High | C45 |
Q8_0Best for your GPU | 8 | 23.5 GB | Very High | C47 |
F16 | 16 | 45.1 GB | Maximum | F0 |
Get started
Copy-paste commands to run Codestral 22B v0.1 IMat on your machine.
Run
lms load hf-legraphista--codestral-22b-v0-1-imat-gguf && lms server startアップグレードオプション
Codestral 22B v0.1 IMatを快適に動かすハードウェア
NVIDIA A100 80GBコスパ重視NVIDIA A800 80GBコスパ最良80 GB VRAM (+32)1935 GB/s (+1263)
C
Raises estimated decode speed by about 225%.112.5 tok/sデコード
Raises estimated decode speed by about 225%.
Adds memory headroom for longer context windows and future model growth.
〜$15,000 MSRP
NVIDIA H800 80GBNVIDIAアップグレード80 GB VRAM (+32)3000 GB/s (+2328)
C
Raises estimated decode speed by about 423%.181.1 tok/sデコード
Raises estimated decode speed by about 423%.
Adds memory headroom for longer context windows and future model growth.
〜$30,000 MSRP
Frequently asked questions
Can Quadro RTX 8000 48GB run Codestral 22B v0.1 IMat?
Yes, Quadro RTX 8000 48GB can run Codestral 22B v0.1 IMat with a C grade (Runs well). Expected decode speed: 34.6 tok/s.
How much VRAM does Codestral 22B v0.1 IMat need?
Codestral 22B v0.1 IMat (22B parameters) requires approximately 22.0 GB of memory with Q4_K_M quantization.
What is the best quantization for Codestral 22B v0.1 IMat?
The recommended quantization for Codestral 22B v0.1 IMat is Q4_K_M, which balances quality and memory efficiency.
What speed will Codestral 22B v0.1 IMat run at on Quadro RTX 8000 48GB?
On Quadro RTX 8000 48GB, Codestral 22B v0.1 IMat achieves approximately 34.6 tokens per second decode speed with a time-to-first-token of 5603ms using Q4_K_M quantization.
Can Quadro RTX 8000 48GB run Codestral 22B v0.1 IMat for coding?
For coding workloads, Codestral 22B v0.1 IMat on Quadro RTX 8000 48GB receives a C grade with 34.6 tok/s and 177K context.
What context window can Codestral 22B v0.1 IMat use on Quadro RTX 8000 48GB?
On Quadro RTX 8000 48GB, Codestral 22B v0.1 IMat can safely use up to 177K tokens of context. The model's official context limit is —, but available memory constrains the safe maximum.
Embed this result▼
Paste this snippet into any page to show a live fit card.
<iframe src="https://willitrunai.com/embed/hf-legraphista--codestral-22b-v0-1-imat-gguf-on-quadro-rtx-8000-48gb" width="400" height="180" frameborder="0" style="border:none;border-radius:12px;overflow:hidden;" title="Will It Run AI — fit result"></iframe>
Preview: