Speed test matrix

Agents-A1-4B-FP8 on RTX 3090 24GB

vllm · 0.32.0 · FP86 contexts (0–248K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 64K, 128K

by @Sandbergllm-inference-bench v0.8.0Measured Oct 30, 2026, 08:08 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: 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), 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 InternScience/Agents-A1-4B-FP8 --max-model-len 262144 --gpu-memory-utilization 0.94 --max-num-seqs 128 --enable-prefix-caching --trust-remote-code --language-model-only

Prefill

prompt tok/s by context length

8K context11.7K721 ms TTFT

16K context11K1,502 ms TTFT

32K context9,9043,285 ms TTFT

64K context8,2637,825 ms TTFT

128K context6,17620,877 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	119.258 ms TTFT	228.970 ms TTFT	438.8134 ms TTFT	799.2198 ms TTFT	1,393.1346 ms TTFT	2,155.8510 ms TTFT	Over capacity	Over capacity
16K	111.270 ms TTFT	203.994 ms TTFT	384.3136 ms TTFT	660.4221 ms TTFT	1,008.1388 ms TTFT	Over capacity	Over capacity	Over capacity
32K	104.2138 ms TTFT	183.5180 ms TTFT	339.5258 ms TTFT	548.3409 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity
64K	92.8210 ms TTFT	152.7277 ms TTFT	273.5420 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
128K	76.0397 ms TTFT	114.4521 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
248K	56.7844 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	11.7K	—	721 ms	—	—	1	—	—
Prefill	16K	1	Valid	11K	—	1,502 ms	—	—	1	—	—
Prefill	32K	1	Valid	9,904.0	—	3,285 ms	—	—	1	—	—
Prefill	64K	1	Valid	8,263.0	—	7,825 ms	—	—	1	—	—
Prefill	128K	1	Valid	6,176.0	—	20,877 ms	—	—	1	—	—
Sustained decode	0	1	Valid	119.2	119.5	58 ms	8.4 ms	1.0	0	3,577	30.0 s
Sustained decode	0	2	Valid	228.9	114.8	70 ms	8.7 ms	2.0	0	6,864	30.0 s
Sustained decode	0	4	Valid	438.8	110.3	134 ms	9.1 ms	4.0	0	13,164	30.0 s
Sustained decode	0	8	Valid	799.2	100.7	198 ms	9.9 ms	8.0	0	23,976	30.0 s
Sustained decode	0	16	Valid	1,393.1	88.0	346 ms	11.4 ms	16.0	0	41,792	30.0 s
Sustained decode	0	32	Valid	2,155.8	67.7	510 ms	14.8 ms	32.0	0	64,672	30.0 s
Sustained decode	0	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	1	Valid	111.2	111.2	70 ms	9.0 ms	1.0	0	3,334	30.0 s
Sustained decode	16K	2	Valid	203.9	102.1	94 ms	9.8 ms	2.0	0	6,116	30.0 s
Sustained decode	16K	4	Valid	384.3	96.3	136 ms	10.4 ms	4.0	0	11,528	30.0 s
Sustained decode	16K	8	Valid	660.4	82.6	221 ms	12.1 ms	8.0	0	19,811	30.0 s
Sustained decode	16K	16	Valid	1,008.1	62.9	388 ms	15.9 ms	16.0	0	30,240	30.0 s
Sustained decode	16K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	1	Valid	104.2	104.4	138 ms	9.6 ms	1.0	0	3,127	30.0 s
Sustained decode	32K	2	Valid	183.5	91.8	180 ms	10.9 ms	2.0	0	5,504	30.0 s
Sustained decode	32K	4	Valid	339.5	84.9	258 ms	11.8 ms	4.0	0	10,184	30.0 s
Sustained decode	32K	8	Valid	548.3	68.2	409 ms	14.7 ms	8.0	0	16,448	30.0 s
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	Valid	92.8	93.0	210 ms	10.8 ms	1.0	0	2,784	30.0 s
Sustained decode	64K	2	Valid	152.7	76.4	277 ms	13.1 ms	2.0	0	4,585	30.0 s
Sustained decode	64K	4	Valid	273.5	68.3	420 ms	14.6 ms	4.0	0	8,208	30.0 s
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	128K	1	Valid	76.0	76.2	397 ms	13.1 ms	1.0	0	2,281	30.0 s
Sustained decode	128K	2	Valid	114.4	57.3	521 ms	17.5 ms	2.0	0	3,432	30.0 s
Sustained decode	128K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	128K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	128K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	128K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	128K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	128K	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	248K	1	Valid	56.7	57.0	844 ms	17.6 ms	1.0	0	1,702	30.0 s
Sustained decode	248K	2	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	248K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	248K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	248K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	248K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	248K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	248K	128	Over capacity	—	—	—	—	—	—	—	—

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