LLM Inference Benchmark Explorer

Gemma-4-26B-A4B-it on 1× DGX Spark — FP16, vLLM, TP1, speculative decoding inference benchmark

Open this configuration in the LLM Inference Benchmark Explorer

ModelGemma-4-26B-A4B-it
Parameters26B
Intelligence Index16.7
Agentic Index2.9
Device1× DGX Spark
QuantizationFP16
Max C4
Chat Capacity16
Agentic Capacity6
TP—
DP—
PP—
EnginevLLM
Speculative DecodingYes
CTTFT (ms)TPS (tok/s)Status
125632.00PASS
266825.20PASS
444820.33PASS
851016.13FAIL
1670712.89FAIL
3274410.63FAIL
649018.69FAIL

In this measurement, Gemma-4-26B-A4B-it (26B parameters), served in FP16 format with vLLM and speculative decoding on 1× DGX Spark, reached a generation speed of 32 tok/s per request and a time to first token (TTFT) of 256 ms with a single request.

Considering the speed targets and the available KV cache capacity, the estimated capacity is 16 users for chat use and 6 for agentic use. The realistic capacity will likely fall between these two values. In scenarios dominated by coding, tool use, long workflows and multi-agent use, capacity approaches the agentic estimate; where shorter interactions, standard conversations and lighter tasks dominate, it approaches the chat estimate.

Intelligence Index and Agentic Index values are published by Artificial Analysis and are reproduced here with attribution.