VLM Inference Benchmark Explorer
Cosmos-Reason2-8B on RTX PRO 6000 — BF16, vLLM vision-language inference benchmark
Open this configuration in the VLM Inference Benchmark Explorer
| Model | Cosmos-Reason2-8B |
|---|---|
| Parameters | 8.8B |
| Device | RTX PRO 6000 |
| Quantization | BF16 |
| Image size | 720p |
| Images per camera | 1 |
| Max Cameras (TTFT ≤ 3000 ms) | 16+ |
| Engine | vLLM |
| Cameras | TTFT (ms) | TPS (tok/s) | Target |
|---|---|---|---|
| 1 | 148 | 92.5 | ≤3000 ms |
| 2 | 203 | 89.2 | ≤3000 ms |
| 4 | 258 | 80.1 | ≤3000 ms |
| 8 | 294 | 67.6 | ≤3000 ms |
| 16 | 545 | 54.8 | ≤3000 ms |
| Images per request | TTFT (ms) | TPS (tok/s) | Target |
|---|---|---|---|
| 1 | 148 | 92.5 | ≤3000 ms |
| 3 | 446 | 92.3 | ≤3000 ms |
| 5 | 750 | 92.4 | ≤3000 ms |
Cosmos-Reason2-8B (8.8B parameters), served in BF16 format with vLLM on RTX PRO 6000, starts answering a single camera that sends one 720p image after 148 ms (time to first token, TTFT), then writes 92.5 tok/s.
With 5 images per camera at 720p, the answer starts after 750 ms. With one 2K image, after 680 ms.
At the default target of a mean TTFT within 3,000 ms (720p, one image per camera), it keeps up with at least 16 cameras at once: it still met the target at the highest number measured, so the real limit is higher.
Cosmos-Reason2-8B (8.8B parameters), served in BF16 format with vLLM on RTX PRO 6000, starts answering a single camera that sends one 720p image after 148 ms (time to first token, TTFT), then writes 92.5 tok/s.
With 5 images per camera at 720p, the answer starts after 750 ms. With one 2K image, after 680 ms.
At the default target of a mean TTFT within 3,000 ms (720p, one image per camera), it keeps up with at least 16 cameras at once: it still met the target at the highest number measured, so the real limit is higher.