Speed test matrix

Nemotron-Cascade-2-30B-A3B-AWQ-4bit on RTX 3090 24GB

vllm · 0.32.0 · AWQ 4-bit6 contexts (0–176K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 64K, 128K

by @Sandbergllm-inference-bench v0.8.0Measured Oct 29, 2026, 21:00 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 4-bit
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 cyankiwi/Nemotron-Cascade-2-30B-A3B-AWQ-4bit --max-model-len 188416 --gpu-memory-utilization 0.94 --max-num-seqs 104 --enable-prefix-caching --trust-remote-code

Prefill

prompt tok/s by context length

8K context9,672870 ms TTFT

16K context9,7211,711 ms TTFT

32K context9,0093,674 ms TTFT

64K context7,8418,417 ms TTFT

128K context6,29220,952 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	151.435 ms TTFT	267.742 ms TTFT	438.6314 ms TTFT	691.4110 ms TTFT	993.9198 ms TTFT	Over capacity	Over capacity	Over capacity
16K	148.1244 ms TTFT	260.4388 ms TTFT	425.2595 ms TTFT	659.21,054 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity
32K	145.4274 ms TTFT	253.1404 ms TTFT	416.7648 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
64K	139.4366 ms TTFT	240.9440 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
128K	131.1689 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
176K	125.8697 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	9,672.0	—	870 ms	—	—	1	—	—
Prefill	16K	1	Valid	9,721.0	—	1,711 ms	—	—	1	—	—
Prefill	32K	1	Valid	9,009.0	—	3,674 ms	—	—	1	—	—
Prefill	64K	1	Valid	7,841.0	—	8,417 ms	—	—	1	—	—
Prefill	128K	1	Valid	6,292.0	—	20,952 ms	—	—	1	—	—
Sustained decode	0	1	Valid	151.4	151.4	35 ms	6.6 ms	1.0	0	4,543	30.0 s
Sustained decode	0	2	Valid	267.7	134.0	42 ms	7.5 ms	2.0	0	8,030	30.0 s
Sustained decode	0	4	Valid	438.6	110.0	314 ms	9.1 ms	4.0	0	13,156	30.0 s
Sustained decode	0	8	Valid	691.4	86.8	110 ms	11.5 ms	8.0	0	20,736	30.0 s
Sustained decode	0	16	Valid	993.9	63.0	198 ms	15.9 ms	16.0	0	29,817	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	148.1	148.2	244 ms	6.7 ms	1.0	0	4,444	30.0 s
Sustained decode	16K	2	Valid	260.4	129.9	388 ms	7.7 ms	2.0	0	7,810	30.0 s
Sustained decode	16K	4	Valid	425.2	105.4	595 ms	9.5 ms	4.0	0	12,756	30.0 s
Sustained decode	16K	8	Valid	659.2	80.8	1,054 ms	12.4 ms	8.0	0	19,776	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	145.4	145.5	274 ms	6.9 ms	1.0	0	4,360	30.0 s
Sustained decode	32K	2	Valid	253.1	126.2	404 ms	7.9 ms	2.0	0	7,591	30.0 s
Sustained decode	32K	4	Valid	416.7	103.3	648 ms	9.7 ms	4.0	0	12,502	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	Valid	139.4	139.8	366 ms	7.2 ms	1.0	0	4,183	30.0 s
Sustained decode	64K	2	Valid	240.9	120.3	440 ms	8.3 ms	2.0	0	7,228	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	128K	1	Valid	131.1	131.2	689 ms	7.6 ms	1.0	0	3,932	30.0 s
Sustained decode	128K	2	Over capacity	—	—	—	—	—	—	—	—
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	176K	1	Valid	125.8	126.2	697 ms	7.9 ms	1.0	0	3,775	30.0 s
Sustained decode	176K	2	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	176K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	176K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	176K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	176K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	176K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	176K	128	Over capacity	—	—	—	—	—	—	—	—

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