Speed test matrix

GLM-5.3 on H200 SXM 141GB ×8

vllm · 0.29.0 · FP86 contexts (8K–251K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 64K, 128K, 251K, 256K

by @keennaylmx dev importMeasured Sep 14, 2026, 18:24 UTC

Hardware

Type: Discrete GPU
GPU: H200 SXM
GPU count: 8
VRAM: 1128 GB (8 × 141 GB)
OS: linux

Engine

Engine: vllm
Version: 0.29.0
Quantization: FP8
Backend: cuda
Repository: github.com/vllm-project/vllm
Build: cuda-13.2

Run

Served model: glm
Revision: main
Benchmark tool: llm-inference-bench v0.6.2
Decode test: Fixed duration
Prefill mode: Measured during decode warm-up
Time per measurement: 30 s
Max output: 4,096 tokens

Launch command

"$HOME/env_h200-glm53-vllm-v0290/bin/vllm" serve zai-org/GLM-5.3 --served-model-name glm --tensor-parallel-size 8 --reasoning-parser glm45 --enable-auto-tool-choice --tool-call-parser glm47 --max-model-len 262144 --gpu-memory-utilization 0.85 --host 0.0.0.0 --port 8000 --api-key <redacted> --dtype bfloat16 --kv-cache-dtype bfloat16 --no-enable-flashinfer-autotune

Prefill

prompt tok/s by context length

8K context12.1K677 ms TTFT

16K context7,6942,109 ms TTFT

32K context11.5K2,806 ms TTFT

64K context10.8K5,971 ms TTFT

128K context9,69813,290 ms TTFT

251K context7,93731,826 ms TTFT

256K context7,91632,544 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	98.449 ms TTFT	185.369 ms TTFT	331.589 ms TTFT	544.4129 ms TTFT	914.6208 ms TTFT	Over capacity	Over capacity	Over capacity
16K	98.472 ms TTFT	185.299 ms TTFT	330.9144 ms TTFT	548.4201 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity
32K	98.2109 ms TTFT	184.8140 ms TTFT	328.9189 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
64K	97.9190 ms TTFT	183.6248 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
128K	96.9366 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity	Over capacity
251K	95.3798 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	12.1K	—	677 ms	—	—	1	—	—
Prefill	16K	1	Valid	7,694.0	—	2,109 ms	—	—	1	—	—
Prefill	32K	1	Valid	11.5K	—	2,806 ms	—	—	1	—	—
Prefill	64K	1	Valid	10.8K	—	5,971 ms	—	—	1	—	—
Prefill	128K	1	Valid	9,698.0	—	13,290 ms	—	—	1	—	—
Prefill	251K	1	Valid	7,937.0	—	31,826 ms	—	—	1	—	—
Prefill	256K	1	Valid	7,916.0	—	32,544 ms	—	—	1	—	—
Sustained decode	8K	1	Valid	98.4	98.3	49 ms	10.2 ms	1.0	0	2,950	30.0 s
Sustained decode	8K	2	Valid	185.3	92.6	69 ms	10.8 ms	2.0	0	5,558	30.0 s
Sustained decode	8K	4	Valid	331.5	82.9	89 ms	12.1 ms	4.0	0	9,944	30.0 s
Sustained decode	8K	8	Valid	544.4	68.1	129 ms	14.7 ms	8.0	0	16,327	30.0 s
Sustained decode	8K	16	Valid	914.6	57.3	208 ms	17.5 ms	16.0	0	27,423	30.0 s
Sustained decode	8K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	8K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	8K	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	1	Valid	98.4	98.4	72 ms	10.2 ms	1.0	0	2,951	30.0 s
Sustained decode	16K	2	Valid	185.2	92.6	99 ms	10.8 ms	2.0	0	5,556	30.0 s
Sustained decode	16K	4	Valid	330.9	82.7	144 ms	12.1 ms	4.0	0	9,924	30.0 s
Sustained decode	16K	8	Valid	548.4	68.5	201 ms	14.6 ms	8.0	0	16,448	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	98.2	98.3	109 ms	10.2 ms	1.0	0	2,947	30.0 s
Sustained decode	32K	2	Valid	184.8	92.4	140 ms	10.8 ms	2.0	0	5,542	30.0 s
Sustained decode	32K	4	Valid	328.9	82.2	189 ms	12.2 ms	4.0	0	9,867	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	97.9	97.9	190 ms	10.2 ms	1.0	0	2,936	30.0 s
Sustained decode	64K	2	Valid	183.6	91.9	248 ms	10.9 ms	2.0	0	5,508	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	96.9	97.0	366 ms	10.3 ms	1.0	0	2,906	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	251K	1	Valid	95.3	95.6	798 ms	10.5 ms	1.0	0	2,859	30.0 s
Sustained decode	251K	2	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	251K	4	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	251K	8	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	251K	16	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	251K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	251K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	251K	128	Over capacity	—	—	—	—	—	—	—	—

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