Does MiMo-V2.6-Flash IQ2_XS_M-v2.1 fit in the NVIDIA DGX Spark (GB10)'s memory?

Estimate: the IQ2_XS_M-v2.1 file is 98.7 GB and needs about 101.7 GB with an ~8K context; the NVIDIA DGX Spark (GB10) has about 96.0 GB usable, so it does not fully fit (part of it runs on the CPU).

Home › Models › MiMo-V2.6-Flash › NVIDIA DGX Spark (GB10)

MiMo-V2.6-Flash on the NVIDIA DGX Spark (GB10)

MiMo-V2.6-Flash runs at about 32.1 tokens/s on the NVIDIA DGX Spark (GB10) (measured, IQ2_XS_M-v2.1, CUDA, llama-bench).

Best measured generation32.1tok/s

IQ2_XS_M-v2.1 · CUDA

Prompt processing422tok/s

same run

For comparison, our size-scaling estimate for this combination is 30.9 tok/s.

Every run

Quant	Backend	Generate tok/s	Prompt tok/s	Test	Build / by / notes	Source
IQ2_XS_M-v2.1	CUDA	32.1	422	tg32	Single DGX Spark, imatrix IQ2 mix (91.9 GiB), default -ub 512 column; the source says the file runs on stock llama.cpp with mimo2 support. File size 98,692,967,808 bytes as stated by the source.	TrevorJS/MiMo-V2.6-Flash-RL-GGUF model card (Hugging Face)

tg128 = generating 128 tokens; pp512 = reading a 512-token prompt (llama-bench). LocalScore averages several prompt and output lengths, so its numbers run lower than tg128. "decode, llama-server" = single-stream generation speed measured against a running llama-server (no speculative decoding).

Check prices

Amazon Newegg

As an Amazon Associate I earn from qualifying purchases.

No card yet?

Try it on a rented GPU (RunPod)

Affiliate link: if you sign up, RigBench may earn a referral credit. It never affects rankings or numbers.

All results for the NVIDIA DGX Spark (GB10) · MiMo-V2.6-Flash on other hardware

Embed this result

HTML

<a href="https://rigbench.pages.dev/model/mimo-v2-6-flash-rl/on/nvidia-dgx-spark-gb10/"><img src="https://rigbench.pages.dev/badge/mimo-v2-6-flash-rl--nvidia-dgx-spark-gb10.svg" alt="MiMo-V2.6-Flash · NVIDIA DGX Spark (GB10): 32.1 tok/s (IQ2_XS_M-v2.1, measured)"></a>

Markdown (GitHub README, forums)

[![MiMo-V2.6-Flash · NVIDIA DGX Spark (GB10): 32.1 tok/s (IQ2_XS_M-v2.1, measured)](https://rigbench.pages.dev/badge/mimo-v2-6-flash-rl--nvidia-dgx-spark-gb10.svg)](https://rigbench.pages.dev/model/mimo-v2-6-flash-rl/on/nvidia-dgx-spark-gb10/)