Speed test matrix
Olmo-3-7B-Instruct on RTX PRO 6000 Blackwell Server Edition 96GB
vllm · 0.30.1rc1.dev151+g73a78e6f1 · BF164 contexts (8K–59K) × 8 concurrency levels (1–128) · prefill at 8K, 16K, 32K, 59K, 64K
Hardware
- Type
- Discrete GPU
- GPU
- RTX PRO 6000 Blackwell Server Edition
- VRAM
- 96 GB
- OS
- linux
Engine
- Engine
- vllm
- Version
- 0.30.1rc1.dev151+g73a78e6f1
- Quantization
- BF16
- Backend
- cuda
- Repository
- github.com/vllm-project/vllm
- Build
- cuda-13.2
- Commit
73a78e6f1f38
Run
- Served model
- olmo
- 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
"$HOME/env_rtxpro6k-allenai-vllm-73a78e6f1f38/bin/vllm" serve allenai/Olmo-3-7B-Instruct --served-model-name olmo --trust-remote-code --tensor-parallel-size 1 --enable-auto-tool-choice --tool-call-parser olmo3 --max-model-len 65536 --gpu-memory-utilization 0.83 --host 0.0.0.0 --port 8000 --api-key <redacted> --attention-backend FLASH_ATTN
Prefill
prompt tok/s by context length8K context23K233 ms TTFT
16K context22.2K478 ms TTFT
32K context21K1,000 ms TTFT
59K context19.4K1,992 ms TTFT
64K context19.1K2,188 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 | 73.131 ms TTFT | 142.337 ms TTFT | 270.652 ms TTFT | 431.378 ms TTFT | 685.0111 ms TTFT | Over capacity | Over capacity | Over capacity |
| 16K | 68.538 ms TTFT | 134.055 ms TTFT | 256.474 ms TTFT | 396.3100 ms TTFT | Over capacity | Over capacity | Over capacity | Over capacity |
| 32K | 60.956 ms TTFT | 119.473 ms TTFT | 230.9101 ms TTFT | Over capacity | Over capacity | Over capacity | Over capacity | Over capacity |
| 59K | 51.288 ms TTFT | 101.0119 ms TTFT | 197.5176 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 | 23K | — | 233 ms | — | — | 1 | — | — |
| Prefill | 16K | 1 | Valid | 22.2K | — | 478 ms | — | — | 1 | — | — |
| Prefill | 32K | 1 | Valid | 21K | — | 1,000 ms | — | — | 1 | — | — |
| Prefill | 59K | 1 | Valid | 19.4K | — | 1,992 ms | — | — | 1 | — | — |
| Prefill | 64K | 1 | Valid | 19.1K | — | 2,188 ms | — | — | 1 | — | — |
| Sustained decode | 8K | 1 | Valid | 73.1 | 73.2 | 31 ms | 13.7 ms | 1.0 | 0 | 2,193 | 30.0 s |
| Sustained decode | 8K | 2 | Valid | 142.3 | 71.3 | 37 ms | 14.0 ms | 2.0 | 0 | 4,268 | 30.0 s |
| Sustained decode | 8K | 4 | Valid | 270.6 | 68.3 | 52 ms | 14.6 ms | 4.0 | 0 | 8,116 | 30.0 s |
| Sustained decode | 8K | 8 | Valid | 431.3 | 54.6 | 78 ms | 18.3 ms | 8.0 | 0 | 12,938 | 30.0 s |
| Sustained decode | 8K | 16 | Valid | 685.0 | 43.6 | 111 ms | 22.9 ms | 16.0 | 0 | 20,544 | 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 | 68.5 | 68.6 | 38 ms | 14.6 ms | 1.0 | 0 | 2,056 | 30.0 s |
| Sustained decode | 16K | 2 | Valid | 134.0 | 67.1 | 55 ms | 14.9 ms | 2.0 | 0 | 4,018 | 30.0 s |
| Sustained decode | 16K | 4 | Valid | 256.4 | 64.6 | 74 ms | 15.5 ms | 4.0 | 0 | 7,688 | 30.0 s |
| Sustained decode | 16K | 8 | Valid | 396.3 | 50.1 | 100 ms | 20.0 ms | 8.0 | 0 | 11,888 | 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 | 60.9 | 61.0 | 56 ms | 16.4 ms | 1.0 | 0 | 1,827 | 30.0 s |
| Sustained decode | 32K | 2 | Valid | 119.4 | 59.8 | 73 ms | 16.7 ms | 2.0 | 0 | 3,582 | 30.0 s |
| Sustained decode | 32K | 4 | Valid | 230.9 | 58.1 | 101 ms | 17.2 ms | 4.0 | 0 | 6,924 | 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 | 59K | 1 | Valid | 51.2 | 51.3 | 88 ms | 19.5 ms | 1.0 | 0 | 1,537 | 30.0 s |
| Sustained decode | 59K | 2 | Valid | 101.0 | 50.6 | 119 ms | 19.8 ms | 2.0 | 0 | 3,028 | 30.0 s |
| Sustained decode | 59K | 4 | Valid | 197.5 | 49.6 | 176 ms | 20.2 ms | 4.0 | 0 | 5,924 | 30.0 s |
| Sustained decode | 59K | 8 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 59K | 16 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 59K | 32 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 59K | 64 | Over capacity | — | — | — | — | — | — | — | — |
| Sustained decode | 59K | 128 | Over capacity | — | — | — | — | — | — | — | — |
Over capacity: not run because this many requests at this context would exceed the server's KV cache.