Speed test matrix
gemma-4-12B-it-int8-AutoRound on RTX PRO 6000 Blackwell Server Edition 96GB
sglang · 0.5.6.post3.dev5791+gf7d4e44d8 · INT85 contexts (8K–123K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 64K, 123K, 128K
Hardware
- Type
- Discrete GPU
- GPU
- RTX PRO 6000 Blackwell Server Edition
- VRAM
- 96 GB
- OS
- linux
Engine
- Engine
- sglang
- Version
- 0.5.6.post3.dev5791+gf7d4e44d8
- Quantization
- INT8
- Backend
- cuda
- Repository
- github.com/sgl-project/sglang
- Build
- cuda-13.2
- Commit
f7d4e44d82ac
Run
- Served model
- gemma
- 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_rtxpro6k-intel-sglang/bin/sglang" serve --model-path Intel/gemma-4-12B-it-int8-AutoRound --served-model-name gemma --trust-remote-code --tp 1 --reasoning-parser gemma4 --tool-call-parser gemma4 --context-length 131072 --mem-fraction-static 0.79 --enable-metrics --host 0.0.0.0 --port 8000 --api-key <redacted>
Prefill
prompt tok/s by context length8K context8,1631,004 ms TTFT
16K context7,3412,213 ms TTFT
32K context6,0795,320 ms TTFT
64K context4,42514,581 ms TTFT
123K context2,57848,049 ms TTFT
128K context2,49451,695 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 | Looped | Looped | Looped | Looped | Looped | Over capacity | Over capacity | Over capacity |
| 16K | 66.5111 ms TTFT | 133.2147 ms TTFT | Looped | 491.9386 ms TTFT | Over capacity | Over capacity | Over capacity | Over capacity |
| 32K | 56.7195 ms TTFT | Looped | Looped | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity |
| 64K | 43.8370 ms TTFT | Looped | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity |
| 123K | 30.8861 ms TTFT | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity |
Looped: the model kept repeating the same text, so the benchmark discarded this measurement. 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 | 8,163.0 | — | 1,004 ms | — | — | 1 | — | — |
| Prefill | 16K | 1 | Valid | 7,341.0 | — | 2,213 ms | — | — | 1 | — | — |
| Prefill | 32K | 1 | Valid | 6,079.0 | — | 5,320 ms | — | — | 1 | — | — |
| Prefill | 64K | 1 | Valid | 4,425.0 | — | 14,581 ms | — | — | 1 | — | — |
| Prefill | 123K | 1 | Valid | 2,578.0 | — | 48,049 ms | — | — | 1 | — | — |
| Prefill | 128K | 1 | Valid | 2,494.0 | — | 51,695 ms | — | — | 1 | — | — |
| Sustained decode | 8K | 1 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 8K | 2 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 8K | 4 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 8K | 8 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 8K | 16 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 8K | 32 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 8K | 64 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 8K | 128 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 16K | 1 | Valid | 66.5 | 66.5 | 111 ms | 15.0 ms | 1.0 | 0 | 1,994 | 30.0 s |
| Sustained decode | 16K | 2 | Valid | 133.2 | 66.6 | 147 ms | 15.0 ms | 2.0 | 0 | 3,996 | 30.0 s |
| Sustained decode | 16K | 4 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 16K | 8 | Valid | 491.9 | 61.5 | 386 ms | 16.2 ms | 8.0 | 0 | 14,752 | 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 | 56.7 | 56.7 | 195 ms | 17.6 ms | 1.0 | 0 | 1,701 | 30.0 s |
| Sustained decode | 32K | 2 | Looped | — | — | — | — | — | 0 | 0 | — |
| Sustained decode | 32K | 4 | Looped | — | — | — | — | — | 0 | 0 | — |
| 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 | 43.8 | 43.8 | 370 ms | 22.8 ms | 1.0 | 0 | 1,313 | 30.0 s |
| Sustained decode | 64K | 2 | Looped | — | — | — | — | — | 0 | 0 | — |
| 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 | 123K | 1 | Valid | 30.8 | 30.8 | 861 ms | 32.4 ms | 1.0 | 0 | 923 | 30.0 s |
| Sustained decode | 123K | 2 | Over capacity | — | — | — | — | — | — | — | — |
| 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 | — | — | — | — | — | — | — | — |
Looped: the model kept repeating the same text, so the benchmark discarded this measurement. Over capacity: not run because this many requests at this context would exceed the server's KV cache.