RTX PRO 6000 Blackwell vs Tesla V100 x2
Best sourced output tok/s per model (single-stream generation) — engine and version, quantization, context and a source link on every row. Hover a value for the quant.
RTX PRO 6000 Blackwell29 sourced runs · best 1271.3 tok/s
Tesla V100 x25 verified runs · best 96.1 tok/s
Specs and evidence
| RTX PRO 6000 Blackwell | Tesla V100 x2 | |
|---|---|---|
| VRAM | 96 GB | 32 GB |
| Sourced runs on the board | 29 | 5 |
| Models measured | 23 | 3 |
| Best output tok/s | 1271.3 tok/s | 96.1 tok/s |
| Highest measured peak VRAM | 72.7 GB | 25.2 GB |
| Street price (latest snapshot) | — | — |
Head-to-head
Only one model has been measured on both cards (Qwen3.8-27B). On that single data point RTX PRO 6000 Blackwell records 296.0 tok/s and Tesla V100 x2 records 24.0 tok/s — treat it as one data point, not a ranking.
The measured matrix
| Model | RTX PRO 6000 Blackwell | Tesla V100 x2 |
|---|---|---|
| MobileLLM-125M | 1271.3 | — |
| Qwen2.5-0.5B-Instruct | 1060.0 | — |
| MobileLLM-350M | 980.8 | — |
| SmolLM2-360M-Instruct | 968.1 | — |
| Qwen3-0.6B | 956.4 | — |
| Llama-3.2-1B-Instruct | 699.9 | — |
| gemma-3-1b-it | 664.3 | — |
| Llama-3.2-3B-Instruct | 489.2 | — |
| Ling-3.0-tiny | 369.1 | — |
| Qwen3.6-35B-A3B | 330.2 | — |
| Ornith-1.0-35B | 309.7 | — |
| Qwen3.8-27B | 296.0 | 24.0 |
| DeepSeek-Coder-V2-Lite-Instruct | 237.3 | — |
| Qwen3-Coder-Next | 174.3 | — |
| Llama-3.1-8B-Instruct | 158.1 | — |
| Muse-Glimmer-30B | 152.2 | — |
| Qwen3.8-27B-MTP | 112.1 | — |
| Qwen3.6-27B | 101.4 | — |
| Qwen3.8-Flash-Next | — | 96.1 |
| Qwen3-14B | 90.8 | — |
| Kimi-Linear-48B-A3B-Instruct | 90.0 | — |
| Ornith-1.5-35B-A3B | — | 65.0 |
| Qwen3.6-27B-MTP | 60.1 | — |
| Qwen2.5-Coder-14B-Instruct | 50.3 | — |
| Qwen3-Coder-30B-A3B-Instruct | 26.1 | — |