Gemma 3 4B IT (QAT) › NVIDIA DGX Spark (GB10)

Gemma 3 4B IT (QAT) on the NVIDIA DGX Spark (GB10)

Best measured: 81.0 tokens/s generation and 5,949 tokens/s prompt processing (Q4_0, CUDA, FA on).

For comparison, our size-scaling estimate for this combination is 94.3 tok/s.

Every run

Quant	Backend	Generate tok/s	Prompt tok/s	Test	Build / by / notes	Source
Q4_0	CUDA FA on	81.0	5,949	tg32	ggml-org maintainers Benchmarked repo: https://huggingface.co/ggml-org/gemma-3-4b-it-qat-GGUF	llama.cpp benches/ (maintainer runs)

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.

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) · Gemma 3 4B IT (QAT) on other hardware