Speed test matrix

Qwen3.8-27B-INT4 on RTX 3090 24GB

vllm · 0.32.0 · INT4 W4A16 g1284 contexts (0–40K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 40K

by @Sandbergllm-inference-bench v0.8.0Measured Oct 29, 2026, 18:48 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: INT4 W4A16 g128
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: MTP, 2 tokens lookahead. KV cache: int4_per_token_head
# 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 RedHatAI/Qwen3.8-27B-INT4 --max-model-len 49152 --gpu-memory-utilization 0.94 --max-num-seqs 31 --enable-prefix-caching --trust-remote-code --language-model-only --kv-cache-dtype int4_per_token_head --attention-backend TRITON_ATTN --speculative-config {"method":"mtp","num_speculative_tokens":2} --max-cudagraph-capture-size 16

Prefill

prompt tok/s by context length

8K context1,7494,708 ms TTFT

16K context1,45511,181 ms TTFT

32K context1,10729,206 ms TTFT

40K context986.040,949 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	59.395 ms TTFT	120.9455 ms TTFT	52.0Requests queued249 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
16K	31.73,395 ms TTFT	63.86,322 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
32K	22.75,893 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
40K	18.65,157 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity

Requests queued: the server could not run all requests at once, so some waited in a queue. 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	1,749.0	—	4,708 ms	—	—	1	—	—
Prefill	16K	1	Valid	1,455.0	—	11,181 ms	—	—	1	—	—
Prefill	32K	1	Valid	1,107.0	—	29,206 ms	—	—	1	—	—
Prefill	40K	1	Valid	986.0	—	40,949 ms	—	—	1	—	—
Sustained decode	0	1	Valid	59.3	60.2	95 ms	16.6 ms	1.0	0	1,779	30.0 s
Sustained decode	0	2	Valid	120.9	61.6	455 ms	16.3 ms	2.0	0	3,626	30.0 s
Sustained decode	0	4	Requests queued	52.0	41.3	249 ms	24.2 ms	2.3	0	1,560	30.0 s
Sustained decode	0	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	1	Valid	31.7	32.6	3,395 ms	30.7 ms	1.0	0	949	29.9 s
Sustained decode	16K	2	Valid	63.8	31.9	6,322 ms	31.4 ms	2.0	0	1,914	30.0 s
Sustained decode	16K	4	Over capacity	—	—	—	—	—	—	—	—
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	22.7	23.4	5,893 ms	42.7 ms	1.0	0	680	30.0 s
Sustained decode	32K	2	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	4	Over capacity	—	—	—	—	—	—	—	—
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	—	—	—	—	—	—	—	—
Sustained decode	40K	1	Valid	18.6	19.4	5,157 ms	51.6 ms	1.0	0	557	29.9 s
Sustained decode	40K	2	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	40K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	40K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	40K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	40K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	40K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	40K	128	Over capacity	—	—	—	—	—	—	—	—

Requests queued: the server could not run all requests at once, so some waited in a queue. Over capacity: not run because this many requests at this context would exceed the server's KV cache.