Speed test matrix

Qwen3-1.7B on RTX 3090 24GB

vllm · 0.32.0 · BF163 contexts (0–32K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K

by @Sandbergllm-inference-bench v0.8.0Measured Oct 31, 2026, 05:25 UTC

Hardware

Type: Discrete GPU
GPU: RTX 3090
VRAM: 24 GB
CPU: Intel Core i7-9700K (8C)
RAM: 32 GB
OS: Ubuntu 24.04.5

Engine

Engine: vllm
Version: 0.32.0
Quantization: BF16
Backend: cuda
Repository: github.com/vllm-project/vllm
Build: vllm/vllm-openai:v0.32.0-x86_64

Run

Served model: bench
Revision: main
Benchmark tool: llm-inference-bench v0.8.0
Decode test: Fixed duration
Prefill mode: Measured during decode warm-up
Time per measurement: 30 s
Max output: 8,192 tokens

Launch command

# Untuned out-of-the-box baseline. Spec decode: none
# 1x RTX 3090 24GB (Zagitta/gpu-unlock), 420 W limit, PCIe Gen3 x16 | Intel Core i7-9700K 8C | 32 GB DDR4 | Ubuntu 24.04.5, NVIDIA driver 580.178.04
vllm serve Qwen/Qwen3-1.7B --max-model-len 40960 --gpu-memory-utilization 0.94 --max-num-seqs 128 --enable-prefix-caching --trust-remote-code

Prefill

prompt tok/s by context length

8K context23.8K347 ms TTFT

16K context19.9K821 ms TTFT

32K context14.3K2,271 ms TTFT

Sustained decode

total tok/s · rows are context length, columns are simultaneous requests

Context	1	2	4	8	16	32	64	128
0	174.014 ms TTFT	314.621 ms TTFT	564.324 ms TTFT	946.130 ms TTFT	1,509.446 ms TTFT	Over capacity	Over capacity	Over capacity
16K	129.558 ms TTFT	200.782 ms TTFT	342.3126 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
32K	102.3106 ms TTFT	145.8140 ms TTFT	235.6211 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity

Over capacity: not run because this many requests at this context would exceed the server's KV cache.

TTFT is measured after the benchmark warms the server's prompt cache, so it mostly excludes prefill.

All measurements

Latencies are medians. Total tok/s is output across all simultaneous requests.

Phase	Context	Requests at once	Status	Total tok/s	Per request tok/s	TTFT	Time between tokens	Average active requests	Completed requests	Output tokens	Measured for
Prefill	8K	1	Valid	23.8K	—	347 ms	—	—	1	—	—
Prefill	16K	1	Valid	19.9K	—	821 ms	—	—	1	—	—
Prefill	32K	1	Valid	14.3K	—	2,271 ms	—	—	1	—	—
Sustained decode	0	1	Valid	174.0	176.6	14 ms	5.7 ms	1.0	0	5,221	30.0 s
Sustained decode	0	2	Valid	314.6	160.9	21 ms	6.2 ms	2.0	0	9,436	30.0 s
Sustained decode	0	4	Valid	564.3	146.3	24 ms	6.8 ms	4.0	0	16,928	30.0 s
Sustained decode	0	8	Valid	946.1	125.4	30 ms	8.0 ms	8.0	0	28,376	30.0 s
Sustained decode	0	16	Valid	1,509.4	102.2	46 ms	9.8 ms	16.0	0	45,280	30.0 s
Sustained decode	0	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	1	Valid	129.5	130.1	58 ms	7.7 ms	1.0	0	3,885	30.0 s
Sustained decode	16K	2	Valid	200.7	101.0	82 ms	9.9 ms	2.0	0	6,021	30.0 s
Sustained decode	16K	4	Valid	342.3	86.2	126 ms	11.6 ms	4.0	0	10,268	30.0 s
Sustained decode	16K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	1	Valid	102.3	102.6	106 ms	9.7 ms	1.0	0	3,069	30.0 s
Sustained decode	32K	2	Valid	145.8	73.1	140 ms	13.7 ms	2.0	0	4,372	30.0 s
Sustained decode	32K	4	Valid	235.6	59.0	211 ms	17.0 ms	4.0	0	7,064	30.0 s
Sustained decode	32K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	128	Over capacity	—	—	—	—	—	—	—	—

Over capacity: not run because this many requests at this context would exceed the server's KV cache.