selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

Agnes-AI logo

Can the NVIDIA DGX Spark (GB10, 128 GB) run Agnes-2.5-Pro-Theta?

Agnes-2.5-Pro-Theta·188B·NVIDIA DGX Spark (GB10, 128 GB)·128 GB unified memory

Saved in this browser, used on every model page.
Just about

Just about. At Q3_K_M it needs about 95 GB of the 96 GB available, so keep the context short, at roughly 1.6 tokens/s.

best quant
Q3_K_M
needs
95 GB
usable
96 GB
est. speed
~1.6 tok/s

Estimates for an 8K context; offload assumes 32 GB of system RAM.Try other contexts on the Agnes-2.5-Pro-Theta page →

Every quantisation of Agnes-2.5-Pro-Theta on the NVIDIA DGX Spark (GB10, 128 GB)

QuantQualityMemoryFitSpeed
Q2_K Heavy loss66 GBfits~2.4 tok/s
Q3_K_M pickNoticeable loss95 GBtight~1.6 tok/s
Q4_K_M Small loss117 GBtoo big—
Q5_K_M Near-lossless136 GBtoo big—
Q6_K Near-lossless157 GBtoo big—
Q8_0 Lossless201 GBtoo big—
FP16 / BF16 Lossless373 GBtoo big—

Speed is a bandwidth-based estimate for generating text (273 GB/s); prompt processing and long contexts are slower.

Measured speed on the NVIDIA DGX Spark (GB10, 128 GB)

No community reports for this pair yet.

Report your own speed →

Run Agnes-2.5-Pro-Theta on the NVIDIA DGX Spark (GB10, 128 GB)

Transformers · Ask about an imagedocs
from transformers import pipeline

pipe = pipeline("image-text-to-text", model="Agnes-AI/Agnes-2.5-Pro-Theta", dtype="auto", device_map="auto")
messages = [
    {
        "role": "user",
        "content": [
            {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/bee.jpg"},
            {"type": "text", "text": "Describe this image."},
        ],
    }
]
out = pipe(text=messages, max_new_tokens=256)
print(out[0]["generated_text"][-1]["content"])

Vision models that fit the NVIDIA DGX Spark (GB10, 128 GB) comfortably

All that fit →
ModelParamsBest quantMemorySpeedLicence
125BA10BQ5_K_M≈91 GBFits in memory~21 tok/sCommercial OK
124BA12BQ5_K_M≈88 GBFits in memory~18 tok/sCommercial OK
109BA18BQ6_K≈88 GBFits in memory~10 tok/sConditions
zai-org logoGLM-4.6Vzai-org
108BA14.3BQ5_K_M≈81 GBFits in memory~15 tok/sCommercial OK
73.4BQ8_0≈79 GBFits in memory~1.9 tok/sConditions
36BA3BQ8_0≈38 GBFits in memory~47 tok/sCommercial OK

Other GPUs that run Agnes-2.5-Pro-Theta

Frequently asked questions

Can the NVIDIA DGX Spark (GB10, 128 GB) run Agnes-2.5-Pro-Theta?
Just about. At Q3_K_M it needs about 95 GB of the 96 GB available, so keep the context short, at roughly 1.6 tokens/s.
How fast does Agnes-2.5-Pro-Theta run on the NVIDIA DGX Spark (GB10, 128 GB)?
An estimate from memory bandwidth gives about 1.6 tokens/s at Q3_K_M.
Which quantisation of Agnes-2.5-Pro-Theta should I use on the NVIDIA DGX Spark (GB10, 128 GB)?
Q3_K_M: the largest one that fits. Expected quality: noticeable loss compared with the full-precision weights.

Memory and speed are estimates from the model's size and the card's memory bandwidth; real numbers depend on the runtime, drivers and context length. Community figures are medians of user reports.