Speed test matrix

Fara1.5-4B-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 31, 2026, 18:48 UTC

Hardware

Type: Discrete GPU
GPU: RTX 3090
VRAM: 24 GB
CPU: AMD Ryzen 9 7900X (12C)
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), 350 W limit (VBIOS max), PCIe Gen4 x16 | AMD Ryzen 9 7900X 12C | 32 GB DDR5 | Ubuntu 24.04.5, NVIDIA driver 595.99.02
vllm serve prithivMLmods/Fara1.5-4B-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 context10.8K760 ms TTFT

16K context10.1K1,607 ms TTFT

32K context9,0433,571 ms TTFT

64K context7,4948,596 ms TTFT

128K context5,57323,093 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	117.324 ms TTFT	226.038 ms TTFT	433.852 ms TTFT	793.7101 ms TTFT	1,389.4142 ms TTFT	2,155.1217 ms TTFT	Over capacity	Over capacity
16K	109.180 ms TTFT	201.4112 ms TTFT	375.8178 ms TTFT	637.4313 ms TTFT	958.7621 ms TTFT	Over capacity	Over capacity	Over capacity
32K	102.579 ms TTFT	180.9100 ms TTFT	328.5144 ms TTFT	519.6238 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity
64K	91.4141 ms TTFT	150.8181 ms TTFT	260.2263 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
128K	75.1362 ms TTFT	113.2440 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
248K	56.2721 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	10.8K	—	760 ms	—	—	1	—	—
Prefill	16K	1	Valid	10.1K	—	1,607 ms	—	—	1	—	—
Prefill	32K	1	Valid	9,043.0	—	3,571 ms	—	—	1	—	—
Prefill	64K	1	Valid	7,494.0	—	8,596 ms	—	—	1	—	—
Prefill	128K	1	Valid	5,573.0	—	23,093 ms	—	—	1	—	—
Sustained decode	0	1	Valid	117.3	117.5	24 ms	8.5 ms	1.0	0	3,518	30.0 s
Sustained decode	0	2	Valid	226.0	113.4	38 ms	8.8 ms	2.0	0	6,780	30.0 s
Sustained decode	0	4	Valid	433.8	109.0	52 ms	9.2 ms	4.0	0	13,012	30.0 s
Sustained decode	0	8	Valid	793.7	100.1	101 ms	10.0 ms	8.0	0	23,808	30.0 s
Sustained decode	0	16	Valid	1,389.4	88.0	142 ms	11.4 ms	16.0	0	41,680	30.0 s
Sustained decode	0	32	Valid	2,155.1	68.4	217 ms	14.6 ms	32.0	0	64,640	30.0 s
Sustained decode	0	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	1	Valid	109.1	109.2	80 ms	9.2 ms	1.0	0	3,273	30.0 s
Sustained decode	16K	2	Valid	201.4	100.8	112 ms	9.9 ms	2.0	0	6,042	30.0 s
Sustained decode	16K	4	Valid	375.8	94.0	178 ms	10.6 ms	4.0	0	11,272	30.0 s
Sustained decode	16K	8	Valid	637.4	79.5	313 ms	12.6 ms	8.0	0	19,120	30.0 s
Sustained decode	16K	16	Valid	958.7	59.4	621 ms	16.8 ms	16.0	0	28,752	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	102.5	102.6	79 ms	9.7 ms	1.0	0	3,075	30.0 s
Sustained decode	32K	2	Valid	180.9	90.6	100 ms	11.0 ms	2.0	0	5,426	30.0 s
Sustained decode	32K	4	Valid	328.5	82.2	144 ms	12.2 ms	4.0	0	9,852	30.0 s
Sustained decode	32K	8	Valid	519.6	64.9	238 ms	15.4 ms	8.0	0	15,584	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	91.4	91.5	141 ms	10.9 ms	1.0	0	2,741	30.0 s
Sustained decode	64K	2	Valid	150.8	75.4	181 ms	13.3 ms	2.0	0	4,522	30.0 s
Sustained decode	64K	4	Valid	260.2	65.1	263 ms	15.4 ms	4.0	0	7,804	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	75.1	75.1	362 ms	13.3 ms	1.0	0	2,251	30.0 s
Sustained decode	128K	2	Valid	113.2	56.5	440 ms	17.7 ms	2.0	0	3,396	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	56.2	56.3	721 ms	17.8 ms	1.0	0	1,685	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.