Speed test matrix

Qwen3.5-27B-AWQ-4bit on RTX 3090 24GB

vllm · 0.32.0 · AWQ 4-bit4 contexts (0–56K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 56K

by @Sandbergllm-inference-bench v0.8.0Measured Oct 29, 2026, 23:36 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: AWQ 4-bit
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. 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 cyankiwi/Qwen3.5-27B-AWQ-4bit --max-model-len 65536 --gpu-memory-utilization 0.94 --max-num-seqs 29 --enable-prefix-caching --trust-remote-code --language-model-only --kv-cache-dtype int4_per_token_head --attention-backend TRITON_ATTN --max-cudagraph-capture-size 16

Prefill

prompt tok/s by context length

8K context1,8044,540 ms TTFT

16K context1,50110,808 ms TTFT

32K context1,14128,309 ms TTFT

56K context839.067,195 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	44.189 ms TTFT	82.0103 ms TTFT	158.4234 ms TTFT	180.5Requests queued351 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity
16K	39.4956 ms TTFT	70.61,391 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
32K	35.62,517 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
56K	28.3Requests queued3,332 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	1,804.0	—	4,540 ms	—	—	1	—	—
Prefill	16K	1	Valid	1,501.0	—	10,808 ms	—	—	1	—	—
Prefill	32K	1	Valid	1,141.0	—	28,309 ms	—	—	1	—	—
Prefill	56K	1	Valid	839.0	—	67,195 ms	—	—	1	—	—
Sustained decode	0	1	Valid	44.1	44.1	89 ms	22.7 ms	1.0	0	1,321	30.0 s
Sustained decode	0	2	Valid	82.0	41.1	103 ms	24.3 ms	2.0	0	2,460	30.0 s
Sustained decode	0	4	Valid	158.4	39.7	234 ms	25.2 ms	4.0	0	4,753	30.0 s
Sustained decode	0	8	Requests queued	180.5	35.0	351 ms	28.6 ms	6.1	0	5,415	30.0 s
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	39.4	39.4	956 ms	25.4 ms	1.0	0	1,181	30.0 s
Sustained decode	16K	2	Valid	70.6	34.9	1,391 ms	28.6 ms	2.0	0	2,116	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	35.6	35.6	2,517 ms	28.1 ms	1.0	0	1,068	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	56K	1	Requests queued	28.3	31.2	3,332 ms	32.1 ms	0.9	0	847	30.0 s
Sustained decode	56K	2	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	56K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	56K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	56K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	56K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	56K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	56K	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.