NVIDIA RTX 4070 Ti Super vs NVIDIA RTX Pro 6000 Blackwell for local LLMs
7 reports on the NVIDIA RTX 4070 Ti Super and 58 on the NVIDIA RTX Pro 6000 Blackwell
Which is faster for local LLMs?
On Qwen3.8 27B at 4-bit, the NVIDIA RTX 4070 Ti Super runs at 5.7 tokens per second (one run) and the NVIDIA RTX Pro 6000 Blackwell at 67.0 (one run). Those are community reports whose engines and settings differ, so the gap is a rough sign, not a controlled measure of the two cards. Memory bandwidth, which caps how fast a card can write, is 2.7x higher on the NVIDIA RTX Pro 6000 Blackwell (672 GB/s on the NVIDIA RTX 4070 Ti Super, 1,792 on the NVIDIA RTX Pro 6000 Blackwell).
- Memory
- 16GB
- Memory bandwidth
- 672 GB/s
- FP16 compute
- 176 TFLOPS
- Reports
- 7
- Memory
- 96GB
- Memory bandwidth
- 1,792 GB/s
- FP16 compute
- 503 TFLOPS
- Reports
- 58
Measured on both
Median tokens per second from plain runs (one device, one request, no speculative decoding, the whole model in the device's memory) of the same model size at the same bit class. Each side shows the engines and reported context behind its median. Context can mean a configured window or prompt depth, so these are community comparisons with differing settings. How to read these
| Model | NVIDIA RTX 4070 Ti Super | NVIDIA RTX Pro 6000 Blackwell |
|---|---|---|
| Qwen3.8 27B4-bit | 5.7one runllama.cpp4K context | 67.0one runengine not statedcontext not stated |
Estimated by the calculator
The quant the calculator recommends for each card and its estimated speed at a 32,768-token context with 32 GB of system RAM, from memory bandwidth and the model's shape. Where a model was also measured above, the measurement wins.
| Model | NVIDIA RTX 4070 Ti Super | NVIDIA RTX Pro 6000 Blackwell |
|---|---|---|
| Qwen3.8 27B | 43.8 t/sQ2_K | 18.3 t/sFP16 |
| Qwen3.8 125B · 6B active | doesn't fit at these settings | 83.1 t/sQ5_K_S |
| DeepSeek V4 Flash 284B · 13B active | doesn't fit at these settings | 130.8 t/sQ2_K |
| Qwen3.6 35B · 3B active | 144.9 t/sQ2_K | 57.5 t/sFP16 |
| Qwen3.6 27B | 43.8 t/sQ2_K | 18.3 t/sFP16 |
| GLM-5.3 320B · 18B active | doesn't fit at these settings | 6.8 t/sQ2_K, experts in RAM |
| Gemma 4 26B · 4B active | 97.7 t/sQ3_K_S | 49.3 t/sFP16 |
Change the context, RAM or model in the calculator: NVIDIA RTX 4070 Ti Super · NVIDIA RTX Pro 6000 Blackwell
Every report on each card, with its source: NVIDIA RTX 4070 Ti Super · NVIDIA RTX Pro 6000 Blackwell