Speed test matrix

Fara1.5-4B on RTX 3090 24GB

vllm · 0.32.0 · BF166 contexts (0–248K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 64K, 128K

by @Sandbergllm-inference-bench v0.8.0Measured Oct 31, 2026, 04:35 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: BF16
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 microsoft/Fara1.5-4B --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 context11.8K695 ms TTFT

16K context11.2K1,452 ms TTFT

32K context9,8643,274 ms TTFT

64K context8,2717,789 ms TTFT

128K context6,22520,675 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	86.535 ms TTFT	169.843 ms TTFT	330.066 ms TTFT	618.4101 ms TTFT	1,121.9179 ms TTFT	1,834.8220 ms TTFT	Over capacity	Over capacity
16K	82.087 ms TTFT	155.4122 ms TTFT	297.2185 ms TTFT	528.7314 ms TTFT	848.8615 ms TTFT	Over capacity	Over capacity	Over capacity
32K	78.2115 ms TTFT	143.2143 ms TTFT	269.8240 ms TTFT	454.5335 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity
64K	71.6209 ms TTFT	123.8252 ms TTFT	226.6386 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
128K	61.2449 ms TTFT	97.2591 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
248K	48.0904 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	11.8K	—	695 ms	—	—	1	—	—
Prefill	16K	1	Valid	11.2K	—	1,452 ms	—	—	1	—	—
Prefill	32K	1	Valid	9,864.0	—	3,274 ms	—	—	1	—	—
Prefill	64K	1	Valid	8,271.0	—	7,789 ms	—	—	1	—	—
Prefill	128K	1	Valid	6,225.0	—	20,675 ms	—	—	1	—	—
Sustained decode	0	1	Valid	86.5	86.6	35 ms	11.6 ms	1.0	0	2,594	30.0 s
Sustained decode	0	2	Valid	169.8	85.1	43 ms	11.8 ms	2.0	0	5,096	30.0 s
Sustained decode	0	4	Valid	330.0	82.8	66 ms	12.1 ms	4.0	0	9,899	30.0 s
Sustained decode	0	8	Valid	618.4	77.7	101 ms	12.9 ms	8.0	0	18,553	30.0 s
Sustained decode	0	16	Valid	1,121.9	70.7	179 ms	14.1 ms	16.0	0	33,648	30.0 s
Sustained decode	0	32	Valid	1,834.8	58.0	220 ms	17.3 ms	32.0	0	55,047	30.0 s
Sustained decode	0	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	1	Valid	82.0	82.0	87 ms	12.2 ms	1.0	0	2,459	30.0 s
Sustained decode	16K	2	Valid	155.4	77.8	122 ms	12.9 ms	2.0	0	4,662	30.0 s
Sustained decode	16K	4	Valid	297.2	74.3	185 ms	13.5 ms	4.0	0	8,912	30.0 s
Sustained decode	16K	8	Valid	528.7	65.9	314 ms	15.2 ms	8.0	0	15,856	30.0 s
Sustained decode	16K	16	Valid	848.8	52.6	615 ms	19.0 ms	16.0	0	25,465	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	78.2	78.3	115 ms	12.8 ms	1.0	0	2,345	30.0 s
Sustained decode	32K	2	Valid	143.2	71.7	143 ms	14.0 ms	2.0	0	4,296	30.0 s
Sustained decode	32K	4	Valid	269.8	67.5	240 ms	14.8 ms	4.0	0	8,092	30.0 s
Sustained decode	32K	8	Valid	454.5	56.7	335 ms	17.6 ms	8.0	0	13,632	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	71.6	71.6	209 ms	14.0 ms	1.0	0	2,147	30.0 s
Sustained decode	64K	2	Valid	123.8	62.0	252 ms	16.1 ms	2.0	0	3,712	30.0 s
Sustained decode	64K	4	Valid	226.6	56.7	386 ms	17.6 ms	4.0	0	6,796	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	61.2	61.3	449 ms	16.3 ms	1.0	0	1,836	30.0 s
Sustained decode	128K	2	Valid	97.2	48.6	591 ms	20.6 ms	2.0	0	2,916	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	48.0	48.2	904 ms	20.7 ms	1.0	0	1,440	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.