Speed test matrix

Muse-Glimmer-30B on RTX PRO 6000 Blackwell Server Edition 96GB

sglang · 0.5.22.dev220+gfa7784d14 · BF165 contexts (8K–123K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 64K, 123K, 128K

by @keennaylmx dev importMeasured Oct 12, 2026, 03:50 UTC

Hardware

Type: Discrete GPU
GPU: RTX PRO 6000 Blackwell Server Edition
VRAM: 96 GB
OS: linux

Engine

Engine: sglang
Version: 0.5.22.dev220+gfa7784d14
Quantization: BF16
Backend: cuda
Repository: github.com/sgl-project/sglang
Build: cuda-13.2
Commit: fa7784d140ee

Run

Served model: muse-glimmer
Revision: main
Benchmark tool: llm-inference-bench v0.7.6
Decode test: Fixed duration
Prefill mode: Measured during decode warm-up
Time per measurement: 30 s
Max output: 4,096 tokens

Launch command

env MAX_JOBS=2 FLASHINFER_NVCC_THREADS=1 TORCHINDUCTOR_COMPILE_THREADS=1 "$HOME/env_rtxpro6k-meta-models-sglang-fa7784d140ee/bin/sglang" serve --model-path meta-models/Muse-Glimmer-30B --served-model-name muse-glimmer --tp 1 --reasoning-parser muse --tool-call-parser muse --context-length 131072 --mem-fraction-static 0.78 --enable-metrics --host 0.0.0.0 --port 8000 --api-key <redacted> --attention-backend flashinfer

Prefill

prompt tok/s by context length

8K context6,7291,218 ms TTFT

16K context6,5552,468 ms TTFT

32K context6,5264,929 ms TTFT

64K context6,14510,439 ms TTFT

123K context5,54622,196 ms TTFT

128K context5,52023,205 ms TTFT

Sustained decode

total tok/s · rows are context length, columns are simultaneous requests

Context	1	2	4	8	16	32	64	128
8K	25.81,297 ms TTFT	51.5215 ms TTFT	102.3257 ms TTFT	202.8301 ms TTFT	403.4464 ms TTFT	766.2603 ms TTFT	Over capacity	Over capacity
16K	25.8189 ms TTFT	51.4292 ms TTFT	102.1308 ms TTFT	202.1394 ms TTFT	400.3640 ms TTFT	Over capacity	Over capacity	Over capacity
32K	25.7216 ms TTFT	51.1297 ms TTFT	101.7418 ms TTFT	200.4585 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity
64K	25.5270 ms TTFT	50.8358 ms TTFT	100.9535 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
123K	25.2355 ms TTFT	50.1493 ms TTFT	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	6,729.0	—	1,218 ms	—	—	1	—	—
Prefill	16K	1	Valid	6,555.0	—	2,468 ms	—	—	1	—	—
Prefill	32K	1	Valid	6,526.0	—	4,929 ms	—	—	1	—	—
Prefill	64K	1	Valid	6,145.0	—	10,439 ms	—	—	1	—	—
Prefill	123K	1	Valid	5,546.0	—	22,196 ms	—	—	1	—	—
Prefill	128K	1	Valid	5,520.0	—	23,205 ms	—	—	1	—	—
Sustained decode	8K	1	Valid	25.8	25.9	1,297 ms	38.6 ms	1.0	0	775	30.0 s
Sustained decode	8K	2	Valid	51.5	25.8	215 ms	38.7 ms	2.0	0	1,544	30.0 s
Sustained decode	8K	4	Valid	102.3	25.7	257 ms	38.9 ms	4.0	0	3,068	30.0 s
Sustained decode	8K	8	Valid	202.8	25.4	301 ms	39.3 ms	8.0	0	6,080	30.0 s
Sustained decode	8K	16	Valid	403.4	25.3	464 ms	39.5 ms	16.0	0	12,096	30.0 s
Sustained decode	8K	32	Valid	766.2	24.0	603 ms	41.6 ms	32.0	0	22,976	30.0 s
Sustained decode	8K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	8K	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	1	Valid	25.8	25.8	189 ms	38.7 ms	1.0	0	773	30.0 s
Sustained decode	16K	2	Valid	51.4	25.8	292 ms	38.8 ms	2.0	0	1,542	30.0 s
Sustained decode	16K	4	Valid	102.1	25.6	308 ms	39.0 ms	4.0	0	3,060	30.0 s
Sustained decode	16K	8	Valid	202.1	25.3	394 ms	39.5 ms	8.0	0	6,063	30.0 s
Sustained decode	16K	16	Valid	400.3	25.1	640 ms	39.8 ms	16.0	0	12,000	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	25.7	25.7	216 ms	38.8 ms	1.0	0	769	30.0 s
Sustained decode	32K	2	Valid	51.1	25.7	297 ms	39.0 ms	2.0	0	1,534	30.0 s
Sustained decode	32K	4	Valid	101.7	25.5	418 ms	39.2 ms	4.0	0	3,052	30.0 s
Sustained decode	32K	8	Valid	200.4	25.2	585 ms	39.8 ms	8.0	0	6,010	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	25.5	25.6	270 ms	39.1 ms	1.0	0	764	30.0 s
Sustained decode	64K	2	Valid	50.8	25.5	358 ms	39.3 ms	2.0	0	1,522	30.0 s
Sustained decode	64K	4	Valid	100.9	25.3	535 ms	39.5 ms	4.0	0	3,025	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	123K	1	Valid	25.2	25.2	355 ms	39.6 ms	1.0	0	754	30.0 s
Sustained decode	123K	2	Valid	50.1	25.1	493 ms	39.8 ms	2.0	0	1,502	30.0 s
Sustained decode	123K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	123K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	123K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	123K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	123K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	123K	128	Over capacity	—	—	—	—	—	—	—	—

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