Speed test matrix

Fara1.5-27B-AWQ-W4A16 on RTX 3090 24GB

vllm · 0.32.0 · AWQ W4A164 contexts (0–56K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 56K

by @Sandbergllm-inference-bench v0.8.0Measured Oct 31, 2026, 07:10 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: AWQ W4A16
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 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 Sohailhosseini/Fara1.5-27B-AWQ-W4A16 --max-model-len 65536 --gpu-memory-utilization 0.94 --max-num-seqs 51 --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 context2,1443,819 ms TTFT

16K context2,0248,013 ms TTFT

32K context1,86317,336 ms TTFT

56K context1,66233,928 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	47.983 ms TTFT	89.098 ms TTFT	171.3227 ms TTFT	332.3368 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity
16K	Looped	85.8519 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
32K	45.1665 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
56K	43.41,242 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity

Looped: the model kept repeating the same text, so the benchmark discarded this measurement. 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	2,144.0	—	3,819 ms	—	—	1	—	—
Prefill	16K	1	Valid	2,024.0	—	8,013 ms	—	—	1	—	—
Prefill	32K	1	Valid	1,863.0	—	17,336 ms	—	—	1	—	—
Prefill	56K	1	Valid	1,662.0	—	33,928 ms	—	—	1	—	—
Sustained decode	0	1	Valid	47.9	48.0	83 ms	20.9 ms	1.0	0	1,437	30.0 s
Sustained decode	0	2	Valid	89.0	44.5	98 ms	22.5 ms	2.0	0	2,668	30.0 s
Sustained decode	0	4	Valid	171.3	42.9	227 ms	23.3 ms	4.0	0	5,140	30.0 s
Sustained decode	0	8	Valid	332.3	41.6	368 ms	24.0 ms	8.0	0	9,969	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	Looped	—	—	—	—	—	0	0	—
Sustained decode	16K	2	Valid	85.8	42.7	519 ms	23.4 ms	2.0	0	2,572	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	45.1	45.1	665 ms	22.2 ms	1.0	0	1,353	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	—	—	—	—	—	—	—	—
Sustained decode	56K	1	Valid	43.4	43.4	1,242 ms	23.1 ms	1.0	0	1,301	30.0 s
Sustained decode	56K	2	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	56K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	56K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	56K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	56K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	56K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	56K	128	Over capacity	—	—	—	—	—	—	—	—

Looped: the model kept repeating the same text, so the benchmark discarded this measurement. Over capacity: not run because this many requests at this context would exceed the server's KV cache.