Can CodeGeeX 4 9B run on Gaudi 3 128GB?

YES — Runs Great

A73Great

Estimated from fit model

CodeGeeX 4 9B needs ~19.8 GB VRAM. Gaudi 3 128GB has 128.0 GB. With Q4_K_M quantization, expect ~126 tok/s.

Runtime: llama.cppCapacity: RoomyBandwidth: HighStack: StandardBottleneck: 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.

Capabilities:

Select quantization to explore

Q4_K_M (Medium quality) — 19.8 GB, 126.0 tok/s, Runs well

19.8 GB required128.0 GB available

15% VRAM used

Fit status

Runs well

Decode

126.0 tok/s

TTFT

1537 ms

Safe context

131K

Memory

19.8 GB / 128.0 GB

Memory breakdown

Weights5.5 GB

KV Cache0.6 GB

Runtime0.9 GB

Headroom12.8 GB

See how fast it feels

See how fast it feelsCodeGeeX 4 9B on Gaudi 3 128GB
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: 126.0 tok/s decode · 1.5s TTFT (warm) · 315 tok/s prefill

What limits this setup

The raw memory story may look fine, but the software ecosystem is still a constraint here.

Runtime ecosystem is narrower than CUDA

Intel GPUs can look attractive on memory per dollar, but local AI tooling, kernels, and model coverage are still broader and easier on CUDA today.

Best improvement path

Prefer CUDA if you want the path of least resistance

If your goal is maximum runtime coverage, easier troubleshooting, and better support for new local AI releases, CUDA is usually still the safer upgrade path.

Performance by workload

Workload	Grade	Fit	Decode	TTFT	Context
Chat	A	Runs well	126.0 tok/s	838 ms	131K
Coding	A	Runs well	126.0 tok/s	1537 ms	131K
Agentic Coding	A	Runs well	126.0 tok/s	2235 ms	131K
Reasoning	A	Runs well	126.0 tok/s	1816 ms	131K
RAG	A	Runs well	126.0 tok/s	2794 ms	131K

Inference speed

CodeGeeX 4 9B inference speed — tokens per second by GPU & Mac

Estimated decode speed (tokens/sec) for CodeGeeX 4 9B at Q4_K_M across popular GPUs and Apple Silicon, using the fastest local runtime per device. Fastest is RTX 5090 32GB at ~126 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	126.0	Fits
RTX 4090 24GB	24 GB	Q4_K_M	126.0	Fits
RTX 3090 24GB	24 GB	Q4_K_M	126.0	Fits
RX 7900 XTX 24GB	24 GB	Q4_K_M	126.0	Fits
RTX 4080 Super 16GB	16 GB	Q4_K_M	121.7	Fits
Mac Studio M3 Ultra 256GB	256 GB	Q4_K_M	111.0	Fits
Mac Studio M2 Ultra 128GB	128 GB	Q4_K_M	92.4	Fits
Mac Studio M1 Ultra 128GB	128 GB	Q4_K_M	87.7	Fits
RTX 4070 12GB	12 GB	Q4_K_M	75.3	Fits
MacBook Pro M4 Max 128GB	128 GB	Q4_K_M	74.7	Fits
MacBook Pro M4 Max 64GB	64 GB	Q4_K_M	74.7	Fits
MacBook Pro M3 Max 64GB	64 GB	Q4_K_M	47.8	Fits
RTX 3060 12GB	12 GB	Q4_K_M	47.3	Fits
MacBook Pro M1 Max 64GB	64 GB	Q4_K_M	43.8	Fits
MacBook Pro M4 Pro 48GB	48 GB	Q4_K_M	38.5	Fits
RTX 4060 8GB	8 GB	Q4_K_M	36.4	Offloads

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 CodeGeeX 4 9B (9B params) fits at each quantization level on Gaudi 3 128GB (128.0 GB usable).

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	3.5 GB	Low	B66
Q3_K_S	3	4.4 GB	Low	B66
NVFP4	4	5.0 GB	Medium	B66
Q4_K_M	4	5.5 GB	Medium	B66
Q5_K_M	5	6.5 GB	High	B66
Q6_K	6	7.4 GB	High	B66
Q8_0	8	9.6 GB	Very High	B66
F16Best for your GPU	16	18.5 GB	Maximum	B67

Get started

Copy-paste commands to run CodeGeeX 4 9B on your machine.

Run

docker run --rm -it ghcr.io/ggerganov/llama.cpp:full \
  --hf-repo "THUDM/codegeex4-all-9b" \
  --hf-file "codegeex4-all-9b-Q4_K_M.gguf" \
  -c 4096 -ngl 99

Your hardware

More models your Gaudi 3 128GB can run

Model	Params	Grade	Decode
Devstral 2 123B Instruct	123B	S	37.5 tok/s
Qwen3-Coder 30B A3B Instruct	30.5B	S	391.6 tok/s
Qwen 3.5 27B	27B	S	169.8 tok/s
Qwen 3.6 27B	27B	S	105.9 tok/s
Qwen 3.5 122B A10B	122B	S	104.1 tok/s

Frequently asked questions

Can Gaudi 3 128GB run CodeGeeX 4 9B?

Yes, Gaudi 3 128GB can run CodeGeeX 4 9B with a A grade (Runs well). Expected decode speed: 126.0 tok/s.

How much VRAM does CodeGeeX 4 9B need?

CodeGeeX 4 9B (9B parameters) requires approximately 19.8 GB of memory with Q4_K_M quantization.

What is the best quantization for CodeGeeX 4 9B?

The recommended quantization for CodeGeeX 4 9B is Q4_K_M, which balances quality and memory efficiency.

What speed will CodeGeeX 4 9B run at on Gaudi 3 128GB?

On Gaudi 3 128GB, CodeGeeX 4 9B achieves approximately 126.0 tokens per second decode speed with a time-to-first-token of 1537ms using Q4_K_M quantization.

Can Gaudi 3 128GB run CodeGeeX 4 9B for coding?

For coding workloads, CodeGeeX 4 9B on Gaudi 3 128GB receives a A grade with 126.0 tok/s and 131K context.

What context window can CodeGeeX 4 9B use on Gaudi 3 128GB?

On Gaudi 3 128GB, CodeGeeX 4 9B can safely use up to 131K tokens of context. The model's official context limit is 131K, but available memory constrains the safe maximum.

What should I upgrade first if CodeGeeX 4 9B feels slow on Gaudi 3 128GB?

Prefer CUDA if you want the path of least resistance. If your goal is maximum runtime coverage, easier troubleshooting, and better support for new local AI releases, CUDA is usually still the safer upgrade path.

Would CUDA be a better path than Gaudi 3 128GB for CodeGeeX 4 9B?

Often yes, if your goal is the easiest setup and the widest runtime support. Intel can offer attractive memory capacity, but CUDA still tends to win on tooling maturity, guides, kernels, and model coverage for local AI.

See all results for Gaudi 3 128GB See all hardware for CodeGeeX 4 9B

Embed this result▼

Paste this snippet into any page to show a live fit card.

<iframe src="https://willitrunai.com/embed/codegeex-4-9b-on-gaudi-3-128gb" width="400" height="180" frameborder="0" style="border:none;border-radius:12px;overflow:hidden;" title="Will It Run AI — fit result"></iframe>

Preview: