Speed test matrix

Qwen3-0.6B-FP8 on RTX 3090 24GB

vllm · 0.32.0 · FP83 contexts (0–32K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K

by @Sandbergllm-inference-bench v0.8.0Measured Oct 30, 2026, 10:22 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: FP8
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/Qwen3-0.6B-FP8 --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 context37.9K218 ms TTFT

16K context28.4K575 ms TTFT

32K context18.4K1,764 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	363.916 ms TTFT	600.724 ms TTFT	811.425 ms TTFT	1,217.632 ms TTFT	1,792.544 ms TTFT	Over capacity	Over capacity	Over capacity
16K	205.656 ms TTFT	287.6130 ms TTFT	459.7112 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
32K	146.8102 ms TTFT	188.9136 ms TTFT	288.2217 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	37.9K	—	218 ms	—	—	1	—	—
Prefill	16K	1	Valid	28.4K	—	575 ms	—	—	1	—	—
Prefill	32K	1	Valid	18.4K	—	1,764 ms	—	—	1	—	—
Sustained decode	0	1	Valid	363.9	380.0	16 ms	2.6 ms	1.0	1	10,917	30.0 s
Sustained decode	0	2	Valid	600.7	336.1	24 ms	3.0 ms	2.0	2	18,020	30.0 s
Sustained decode	0	4	Valid	811.4	225.4	25 ms	4.4 ms	4.0	0	24,340	30.0 s
Sustained decode	0	8	Valid	1,217.6	177.1	32 ms	5.6 ms	8.0	0	36,529	30.0 s
Sustained decode	0	16	Valid	1,792.5	132.5	44 ms	7.5 ms	16.0	0	53,776	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	205.6	207.9	56 ms	4.8 ms	1.0	0	6,169	30.0 s
Sustained decode	16K	2	Valid	287.6	145.8	130 ms	6.9 ms	2.0	0	8,628	30.0 s
Sustained decode	16K	4	Valid	459.7	116.5	112 ms	8.6 ms	4.0	0	13,789	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	146.8	147.6	102 ms	6.8 ms	1.0	0	4,402	30.0 s
Sustained decode	32K	2	Valid	188.9	94.9	136 ms	10.5 ms	2.0	0	5,666	30.0 s
Sustained decode	32K	4	Valid	288.2	72.2	217 ms	13.8 ms	4.0	0	8,644	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.