benchmarks / gemma-4-26b-a4b-it--dgx-spark-x1--llama-cpp--q4-k-m--greedy-baseline--p2048-o256-c1--2026-07-06

Gemma 4 26B-A4B on NVIDIA DGX Spark

compare

50.5tok/s

926msttft

2.6kprefill

llama.cpp b15-ee445f9Q4_K_Mspec-decodingJul 6, 2026

Run configuration

nodes

batch size

—

concurrency

context

10240

prompt / output

2036 / 256

weights unsloth/gemma-4-26B-A4B-it-GGUF ↗

Metrics

decode tok/s

50.5

prefill tok/s

2.6k

ttft

926ms

total tok/s

43.2

inter-token

20ms

e2e latency

5.94s

req/s

0.169

peak memory

25 GB

avg power

37.8 W

energy/req

224.2 J

temperature

Reproduce

~/Dev/llama.cpp/build/bin/llama-server -m ~/models/gguf/gemma-4-26B-A4B-it-UD-Q4_K_M.gguf -c 10240 --parallel 1 -ngl 99 -fa on --host 127.0.0.1 --port 8080
python3 bench/harness.py bench/configs/gemma-4-26b-a4b-it--spec-baseline.json --rows-out data/benchmarks/gemma-4-26b-a4b-it.json

spec-decoding profile: greedy (temp 0) no-draft baseline; only comparable to other spec-decoding rows.