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 31, 2026, 20:14 UTC

Hardware

Type: Discrete GPU
GPU: RTX 3090
VRAM: 24 GB
CPU: AMD Ryzen 9 7900X (12C)
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), 350 W limit (VBIOS max), PCIe Gen4 x16 | AMD Ryzen 9 7900X 12C | 32 GB DDR5 | Ubuntu 24.04.5, NVIDIA driver 595.99.02
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 context36.4K227 ms TTFT

16K context26.8K610 ms TTFT

32K context16.8K1,932 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	357.09 ms TTFT	594.912 ms TTFT	804.817 ms TTFT	1,204.019 ms TTFT	1,763.951 ms TTFT	Over capacity	Over capacity	Over capacity
16K	201.042 ms TTFT	281.453 ms TTFT	423.079 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
32K	143.176 ms TTFT	185.4104 ms TTFT	258.5154 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	36.4K	—	227 ms	—	—	1	—	—
Prefill	16K	1	Valid	26.8K	—	610 ms	—	—	1	—	—
Prefill	32K	1	Valid	16.8K	—	1,932 ms	—	—	1	—	—
Sustained decode	0	1	Valid	357.0	373.1	9 ms	2.7 ms	1.0	1	10,709	30.0 s
Sustained decode	0	2	Valid	594.9	334.9	12 ms	3.0 ms	2.0	2	17,847	30.0 s
Sustained decode	0	4	Valid	804.8	225.3	17 ms	4.4 ms	4.0	0	24,140	30.0 s
Sustained decode	0	8	Valid	1,204.0	174.3	19 ms	5.7 ms	8.0	0	36,112	30.0 s
Sustained decode	0	16	Valid	1,763.9	131.5	51 ms	7.6 ms	16.0	0	52,912	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	201.0	203.1	42 ms	4.9 ms	1.0	0	6,030	30.0 s
Sustained decode	16K	2	Valid	281.4	142.6	53 ms	7.0 ms	2.0	0	8,442	30.0 s
Sustained decode	16K	4	Valid	423.0	107.2	79 ms	9.3 ms	4.0	0	12,688	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	143.1	143.9	76 ms	6.9 ms	1.0	0	4,294	30.0 s
Sustained decode	32K	2	Valid	185.4	93.1	104 ms	10.7 ms	2.0	0	5,560	30.0 s
Sustained decode	32K	4	Valid	258.5	64.8	154 ms	15.4 ms	4.0	0	7,752	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.