Speed test matrix

INTELLECT-3-FP8 on H200 NVL 141GB ×2

vllm · 0.26.1rc1.dev638+g52be12cfa · FP85 contexts (8K–123K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 64K, 123K, 128K

by @keennaylmx dev importMeasured Sep 6, 2026, 02:26 UTC

Hardware

Type: Discrete GPU
GPU: H200 NVL
GPU count: 2
VRAM: 282 GB (2 × 141 GB)
OS: linux

Engine

Engine: vllm
Version: 0.26.1rc1.dev638+g52be12cfa
Quantization: FP8
Backend: cuda
Repository: github.com/vllm-project/vllm
Build: cuda-13.2
Commit: 52be12cfac0c

Run

Served model: intellect
Revision: main
Benchmark tool: llm-inference-bench v0.4.29
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-primeintellect-vllm/bin/vllm" serve PrimeIntellect/INTELLECT-3-FP8 --served-model-name intellect --trust-remote-code --tensor-parallel-size 2 --reasoning-parser deepseek_r1 --enable-auto-tool-choice --tool-call-parser qwen3_coder --max-model-len 131072 --gpu-memory-utilization 0.874440 --host 0.0.0.0 --port 8000 --api-key <redacted>

Prefill

prompt tok/s by context length

8K context24.2K339 ms TTFT

16K context19.9K816 ms TTFT

32K context15.1K2,139 ms TTFT

64K context9,9866,456 ms TTFT

123K context6,08920,328 ms TTFT

128K context5,95121,643 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	132.451 ms TTFT	227.956 ms TTFT	387.789 ms TTFT	627.9134 ms TTFT	967.5202 ms TTFT	1,491.4318 ms TTFT	Over capacity	Over capacity
16K	128.588 ms TTFT	219.4102 ms TTFT	371.0131 ms TTFT	584.7203 ms TTFT	875.7308 ms TTFT	1,293.7432 ms TTFT	Over capacity	Over capacity
32K	122.4102 ms TTFT	206.9123 ms TTFT	342.8209 ms TTFT	515.0272 ms TTFT	736.3394 ms TTFT	Over capacity	Over capacity	Over capacity
64K	112.0170 ms TTFT	185.4225 ms TTFT	295.8360 ms TTFT	413.5502 ms TTFT	Over capacity	Over capacity	Over capacity	Over capacity
123K	97.4276 ms TTFT	156.2358 ms TTFT	236.2493 ms TTFT	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	24.2K	—	339 ms	—	—	1	—	—
Prefill	16K	1	Valid	19.9K	—	816 ms	—	—	1	—	—
Prefill	32K	1	Valid	15.1K	—	2,139 ms	—	—	1	—	—
Prefill	64K	1	Valid	9,986.0	—	6,456 ms	—	—	1	—	—
Prefill	123K	1	Valid	6,089.0	—	20,328 ms	—	—	1	—	—
Prefill	128K	1	Valid	5,951.0	—	21,643 ms	—	—	1	—	—
Sustained decode	8K	1	Valid	132.4	133.5	51 ms	7.5 ms	1.0	1	3,972	30.0 s
Sustained decode	8K	2	Valid	227.9	114.1	56 ms	8.8 ms	2.0	0	6,834	30.0 s
Sustained decode	8K	4	Valid	387.7	97.1	89 ms	10.3 ms	4.0	0	11,628	30.0 s
Sustained decode	8K	8	Valid	627.9	78.8	134 ms	12.7 ms	8.0	0	18,832	30.0 s
Sustained decode	8K	16	Valid	967.5	60.9	202 ms	16.4 ms	16.0	0	29,024	30.0 s
Sustained decode	8K	32	Valid	1,491.4	47.0	318 ms	21.3 ms	32.0	0	44,736	30.0 s
Sustained decode	8K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	8K	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	1	Valid	128.5	129.7	88 ms	7.7 ms	1.0	1	3,854	30.0 s
Sustained decode	16K	2	Valid	219.4	109.9	102 ms	9.1 ms	2.0	0	6,586	30.0 s
Sustained decode	16K	4	Valid	371.0	93.0	131 ms	10.8 ms	4.0	0	11,128	30.0 s
Sustained decode	16K	8	Valid	584.7	73.3	203 ms	13.6 ms	8.0	0	17,536	30.0 s
Sustained decode	16K	16	Valid	875.7	55.0	308 ms	18.2 ms	16.0	0	26,272	30.0 s
Sustained decode	16K	32	Valid	1,293.7	40.7	432 ms	24.6 ms	32.0	0	38,816	30.0 s
Sustained decode	16K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	16K	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	1	Valid	122.4	122.5	102 ms	8.2 ms	1.0	1	3,668	30.0 s
Sustained decode	32K	2	Valid	206.9	103.6	123 ms	9.6 ms	2.0	0	6,208	30.0 s
Sustained decode	32K	4	Valid	342.8	85.9	209 ms	11.6 ms	4.0	0	10,284	30.0 s
Sustained decode	32K	8	Valid	515.0	64.6	272 ms	15.5 ms	8.0	0	15,448	30.0 s
Sustained decode	32K	16	Valid	736.3	46.2	394 ms	21.6 ms	16.0	0	22,080	30.0 s
Sustained decode	32K	32	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	64	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	32K	128	Over capacity	—	—	—	—	—	—	—	—
Sustained decode	64K	1	Valid	112.0	112.3	170 ms	8.9 ms	1.0	0	3,361	30.0 s
Sustained decode	64K	2	Valid	185.4	92.9	225 ms	10.8 ms	2.0	0	5,562	30.0 s
Sustained decode	64K	4	Valid	295.8	74.1	360 ms	13.5 ms	4.0	0	8,872	30.0 s
Sustained decode	64K	8	Valid	413.5	51.8	502 ms	19.3 ms	8.0	0	12,407	30.0 s
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	97.4	97.6	276 ms	10.2 ms	1.0	0	2,923	30.0 s
Sustained decode	123K	2	Valid	156.2	78.2	358 ms	12.8 ms	2.0	0	4,684	30.0 s
Sustained decode	123K	4	Valid	236.2	59.2	493 ms	16.9 ms	4.0	0	7,084	30.0 s
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.