Speed test matrix

Qwen3.5-9B-AWQ-4bit on RTX 3060 12GB

vllm · 0.32.0 · W4A16 g323 contexts (0–32K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K

by @Sandbergllm-inference-bench v0.8.0Measured Nov 1, 2026, 16:17 UTC

Hardware

Type: Discrete GPU
GPU: RTX 3060
VRAM: 12 GB
CPU: Intel Xeon E5-2680 v4 (14C)
RAM: 32 GB
OS: Ubuntu 24.04.5

Engine

Engine: vllm
Version: 0.32.0
Quantization: W4A16 g32
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: fp8
# 1x RTX 3060 12GB, stock 170 W limit, PCIe Gen3 x16 | Intel Xeon E5-2680 v4 14C | 32 GB DDR4-2400 | Ubuntu 24.04.5, NVIDIA driver 580.178.04
vllm serve cyankiwi/Qwen3.5-9B-AWQ-4bit --max-model-len 40960 --gpu-memory-utilization 0.94 --max-num-seqs 74 --enable-prefix-caching --trust-remote-code --language-model-only --kv-cache-dtype fp8 --max-cudagraph-capture-size 16

Prefill

prompt tok/s by context length

8K context1,5935,115 ms TTFT

16K context1,50210,735 ms TTFT

32K context1,37723,306 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	50.1135 ms TTFT	97.8236 ms TTFT	187.7285 ms TTFT	345.3525 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity
16K	48.2330 ms TTFT	91.9487 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
32K	46.4560 ms TTFT	Over capacity	Over capacity	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	1,593.0	—	5,115 ms	—	—	1	—	—
Prefill	16K	1	Valid	1,502.0	—	10,735 ms	—	—	1	—	—
Prefill	32K	1	Valid	1,377.0	—	23,306 ms	—	—	1	—	—
Sustained decode	0	1	Valid	50.1	50.1	135 ms	19.9 ms	1.0	0	1,503	30.0 s
Sustained decode	0	2	Valid	97.8	49.0	236 ms	20.4 ms	2.0	0	2,934	30.0 s
Sustained decode	0	4	Valid	187.7	47.0	285 ms	21.3 ms	4.0	0	5,627	30.0 s
Sustained decode	0	8	Valid	345.3	43.3	525 ms	23.1 ms	8.0	0	10,360	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	48.2	48.2	330 ms	20.7 ms	1.0	0	1,445	30.0 s
Sustained decode	16K	2	Valid	91.9	45.9	487 ms	21.8 ms	2.0	0	2,754	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	46.4	46.5	560 ms	21.5 ms	1.0	0	1,392	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	—	—	—	—	—	—	—	—

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