Speed test matrix

granite-4.2-8b-nvfp4 on RTX 3060 12GB

vllm · 0.32.0 · NVFP45 contexts (0–80K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 64K, 80K

by @Sandbergllm-inference-bench v0.8.0Measured Oct 31, 2026, 12:01 UTC

Hardware

Type: Discrete GPU
GPU: RTX 3060
VRAM: 12 GB
CPU: Intel Xeon E5-2680 v4 (14C)
RAM: 32 GB
OS: Ubuntu 24.04.5

Engine

Engine: vllm
Version: 0.32.0
Quantization: NVFP4
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: int4_per_token_head
# 1x RTX 3060 12GB, stock 170 W limit, PCIe Gen3 x16 | Intel Xeon E5-2680 v4 14C | 32 GB DDR4-2400 | Ubuntu 24.04.5, NVIDIA driver 580.178.04
vllm serve ibm-granite/granite-4.2-8b-nvfp4 --max-model-len 90112 --gpu-memory-utilization 0.94 --max-num-seqs 128 --enable-prefix-caching --trust-remote-code --kv-cache-dtype int4_per_token_head --attention-backend TRITON_ATTN

Prefill

prompt tok/s by context length

8K context910.09,014 ms TTFT

16K context652.024,875 ms TTFT

32K context425.076,012 ms TTFT

64K context251.0256,441 ms TTFT

80K context209.0385,135 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	49.7123 ms TTFT	95.4137 ms TTFT	176.570 ms TTFT	311.694 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity
16K	30.2192 ms TTFT	47.6284 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
32K	21.7Requests queued313 ms TTFT	31.6Requests queued405 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
64K	14.0Requests queued593 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
80K	11.8Requests queued527 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity

Requests queued: the server could not run all requests at once, so some waited in a queue. 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	910.0	—	9,014 ms	—	—	1	—	—
Prefill	16K	1	Valid	652.0	—	24,875 ms	—	—	1	—	—
Prefill	32K	1	Valid	425.0	—	76,012 ms	—	—	1	—	—
Prefill	64K	1	Valid	251.0	—	256,441 ms	—	—	1	—	—
Prefill	80K	1	Valid	209.0	—	385,135 ms	—	—	1	—	—
Sustained decode	0	1	Valid	49.7	50.0	123 ms	20.0 ms	1.0	0	1,491	30.0 s
Sustained decode	0	2	Valid	95.4	48.2	137 ms	20.8 ms	2.0	0	2,863	30.0 s
Sustained decode	0	4	Valid	176.5	44.9	70 ms	22.3 ms	4.0	0	5,292	30.0 s
Sustained decode	0	8	Valid	311.6	40.0	94 ms	25.0 ms	8.0	0	9,344	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	Valid	30.2	30.3	192 ms	33.0 ms	1.0	0	906	30.0 s
Sustained decode	16K	2	Valid	47.6	23.8	284 ms	41.9 ms	2.0	0	1,426	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	Requests queued	21.7	21.8	313 ms	46.0 ms	1.0	0	651	30.0 s
Sustained decode	32K	2	Requests queued	31.6	15.8	405 ms	63.3 ms	2.0	0	948	30.0 s
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	64K	1	Requests queued	14.0	14.0	593 ms	71.3 ms	1.0	0	419	30.0 s
Sustained decode	64K	2	Over capacity	—	—	—	—	—	—	—	—
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	80K	1	Requests queued	11.8	11.9	527 ms	84.2 ms	1.0	0	355	30.0 s
Sustained decode	80K	2	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	80K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	80K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	80K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	80K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	80K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	80K	128	Over capacity	—	—	—	—	—	—	—	—

Requests queued: the server could not run all requests at once, so some waited in a queue. Over capacity: not run because this many requests at this context would exceed the server's KV cache.