Speed test matrix

Agents-A1-4B-FP8 on RTX 3060 12GB

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, 04:14 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: 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. 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 InternScience/Agents-A1-4B-FP8 --max-model-len 262144 --gpu-memory-utilization 0.94 --max-num-seqs 128 --enable-prefix-caching --trust-remote-code --language-model-only --kv-cache-dtype int4_per_token_head --attention-backend TRITON_ATTN

Prefill

prompt tok/s by context length

8K context2,3413,583 ms TTFT

16K context1,9618,341 ms TTFT

32K context1,48921,721 ms TTFT

64K context1,00264,156 ms TTFT

128K context606.0211,378 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	49.6164 ms TTFT	94.5264 ms TTFT	180.3547 ms TTFT	330.2896 ms TTFT	566.41,574 ms TTFT	890.62,368 ms TTFT	Over capacity	Over capacity
16K	44.71,416 ms TTFT	80.12,125 ms TTFT	136.03,482 ms TTFT	213.36,198 ms TTFT	298.211,703 ms TTFT	Over capacity	Over capacity	Over capacity
32K	40.71,766 ms TTFT	69.72,618 ms TTFT	108.94,266 ms TTFT	157.27,567 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity
64K	34.51,698 ms TTFT	55.12,418 ms TTFT	77.93,843 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
128K	23.4Requests queued4,124 ms TTFT	29.5Requests queued6,119 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
248K	Over capacity	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	2,341.0	—	3,583 ms	—	—	1	—	—
Prefill	16K	1	Valid	1,961.0	—	8,341 ms	—	—	1	—	—
Prefill	32K	1	Valid	1,489.0	—	21,721 ms	—	—	1	—	—
Prefill	64K	1	Valid	1,002.0	—	64,156 ms	—	—	1	—	—
Prefill	128K	1	Valid	606.0	—	211,378 ms	—	—	1	—	—
Sustained decode	0	1	Valid	49.6	49.7	164 ms	20.1 ms	1.0	0	1,488	30.0 s
Sustained decode	0	2	Valid	94.5	47.3	264 ms	21.2 ms	2.0	0	2,834	30.0 s
Sustained decode	0	4	Valid	180.3	45.2	547 ms	22.1 ms	4.0	0	5,408	30.0 s
Sustained decode	0	8	Valid	330.2	41.3	896 ms	24.2 ms	8.0	0	9,904	30.0 s
Sustained decode	0	16	Valid	566.4	35.3	1,574 ms	28.3 ms	16.0	0	16,993	30.0 s
Sustained decode	0	32	Valid	890.6	26.9	2,368 ms	37.1 ms	32.0	0	26,688	30.0 s
Sustained decode	0	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	0	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	1	Valid	44.7	44.8	1,416 ms	22.3 ms	1.0	0	1,340	30.0 s
Sustained decode	16K	2	Valid	80.1	39.4	2,125 ms	25.4 ms	2.0	0	2,402	30.0 s
Sustained decode	16K	4	Valid	136.0	32.2	3,482 ms	31.1 ms	4.0	0	4,080	30.0 s
Sustained decode	16K	8	Valid	213.3	23.6	6,198 ms	42.3 ms	8.0	0	6,392	30.0 s
Sustained decode	16K	16	Valid	298.2	14.6	11,703 ms	68.4 ms	16.0	0	8,944	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	40.7	40.8	1,766 ms	24.5 ms	1.0	0	1,220	30.0 s
Sustained decode	32K	2	Valid	69.7	34.1	2,618 ms	29.3 ms	2.0	0	2,090	30.0 s
Sustained decode	32K	4	Valid	108.9	25.5	4,266 ms	39.2 ms	4.0	0	3,268	30.0 s
Sustained decode	32K	8	Valid	157.2	17.0	7,567 ms	59.0 ms	8.0	0	4,712	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	34.5	34.6	1,698 ms	28.9 ms	1.0	0	1,035	30.0 s
Sustained decode	64K	2	Valid	55.1	27.1	2,418 ms	36.9 ms	2.0	0	1,652	30.0 s
Sustained decode	64K	4	Valid	77.9	18.5	3,843 ms	54.1 ms	4.0	0	2,336	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	Requests queued	23.4	26.5	4,124 ms	37.7 ms	0.9	0	701	30.0 s
Sustained decode	128K	2	Requests queued	29.5	18.2	6,119 ms	55.2 ms	1.7	0	884	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	Over capacity	—	—	—	—	0.7	0	0	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	—	—	—	—	—	—	—	—

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.