Speed test matrix

Llama-3.2-3B-Instruct-FP8-dynamic on RTX 3060 12GB

vllm · 0.32.0 · FP8 dynamic5 contexts (0–120K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 64K, 120K

by @Sandbergllm-inference-bench v0.8.0Measured Oct 31, 2026, 16:47 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: FP8 dynamic
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 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 RedHatAI/Llama-3.2-3B-Instruct-FP8-dynamic --max-model-len 131072 --gpu-memory-utilization 0.94 --max-num-seqs 128 --enable-prefix-caching --trust-remote-code  --kv-cache-dtype int4_per_token_head --attention-backend TRITON_ATTN

Prefill

prompt tok/s by context length

8K context1,8404,463 ms TTFT

16K context1,19013,640 ms TTFT

32K context713.045,274 ms TTFT

64K context395.0163,214 ms TTFT

120K context222.0543,621 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	71.597 ms TTFT	134.9156 ms TTFT	243.4159 ms TTFT	416.5175 ms TTFT	664.2209 ms TTFT	Over capacity	Over capacity	Over capacity
16K	43.7153 ms TTFT	63.3204 ms TTFT	86.8279 ms TTFT	108.9472 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity
32K	31.3265 ms TTFT	40.6373 ms TTFT	52.0540 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
64K	19.9Requests queued519 ms TTFT	23.5Requests queued667 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
120K	12.1Requests queued844 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	1,840.0	—	4,463 ms	—	—	1	—	—
Prefill	16K	1	Valid	1,190.0	—	13,640 ms	—	—	1	—	—
Prefill	32K	1	Valid	713.0	—	45,274 ms	—	—	1	—	—
Prefill	64K	1	Valid	395.0	—	163,214 ms	—	—	1	—	—
Prefill	120K	1	Valid	222.0	—	543,621 ms	—	—	1	—	—
Sustained decode	0	1	Valid	71.5	72.2	97 ms	13.9 ms	1.0	0	2,143	30.0 s
Sustained decode	0	2	Valid	134.9	68.5	156 ms	14.6 ms	2.0	0	4,047	30.0 s
Sustained decode	0	4	Valid	243.4	62.4	159 ms	16.0 ms	4.0	0	7,300	30.0 s
Sustained decode	0	8	Valid	416.5	54.1	175 ms	18.5 ms	8.0	0	12,496	30.0 s
Sustained decode	0	16	Valid	664.2	43.9	209 ms	22.8 ms	16.0	0	19,920	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	43.7	43.8	153 ms	22.8 ms	1.0	0	1,311	30.0 s
Sustained decode	16K	2	Valid	63.3	31.7	204 ms	31.5 ms	2.0	0	1,898	30.0 s
Sustained decode	16K	4	Valid	86.8	21.8	279 ms	45.9 ms	4.0	0	2,604	30.0 s
Sustained decode	16K	8	Valid	108.9	13.6	472 ms	73.3 ms	8.0	0	3,264	30.0 s
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	31.3	31.4	265 ms	31.9 ms	1.0	0	938	30.0 s
Sustained decode	32K	2	Valid	40.6	20.3	373 ms	49.2 ms	2.0	0	1,216	30.0 s
Sustained decode	32K	4	Valid	52.0	13.0	540 ms	76.9 ms	4.0	0	1,560	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	—	—	—	—	—	—	—	—
Sustained decode	64K	1	Requests queued	19.9	20.0	519 ms	50.0 ms	1.0	0	598	30.0 s
Sustained decode	64K	2	Requests queued	23.5	11.8	667 ms	84.6 ms	2.0	0	705	30.0 s
Sustained decode	64K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	120K	1	Requests queued	12.1	12.2	844 ms	82.0 ms	1.0	0	362	30.0 s
Sustained decode	120K	2	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	120K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	120K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	120K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	120K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	120K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	120K	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.