M3 Ultra 96GB vs NVIDIA RTX Pro 6000 Blackwell for local LLMs
10 reports on the M3 Ultra 96GB and 70 on the NVIDIA RTX Pro 6000 Blackwell
Which is faster for local LLMs?
On Qwen3.5 35B · 3B active at 8-bit, the M3 Ultra 96GB has 73.4 generation tokens/s (one run); prompt processing 1808.0 tokens/s (one run); the NVIDIA RTX Pro 6000 Blackwell has generation speed not reported; prompt processing 16459.6 tokens/s (one run). Memory bandwidth, which caps how fast a card can write, is 2.2x higher on the NVIDIA RTX Pro 6000 Blackwell (819 GB/s on the M3 Ultra 96GB, 1,792 on the NVIDIA RTX Pro 6000 Blackwell).
- Memory
- 96GB unified
- Memory bandwidth
- 819 GB/s
- FP16 compute
- 21.3 TFLOPS
- Reports
- 10
- Memory
- 96GB
- Memory bandwidth
- 1,792 GB/s
- FP16 compute
- 503 TFLOPS
- Reports
- 70
Measured on both
Generation and prompt processing speeds 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 | M3 Ultra 96GB | NVIDIA RTX Pro 6000 Blackwell |
|---|---|---|
| Qwen3.5 35B · 3B active8-bit | 73.4one runPP 1808.0 (one run)llama.cppcontext not stated | -0 runsPP 16459.6 (one run)llama.cpp2K context |
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 | M3 Ultra 96GB | NVIDIA RTX Pro 6000 Blackwell |
|---|---|---|
| Qwen3.5 35B · 3B active | 57.1 t/sQ8_0 | 57.5 t/sFP16 |
| Qwen3.8 27B | 7.9 t/sFP16 | 18.3 t/sFP16 |
| Qwen3.8 125B · 6B active | 55.1 t/sQ3_K_M | 83.1 t/sQ5_K_S |
| Qwen3.6 35B · 3B active | 57.1 t/sQ8_0 | 57.5 t/sFP16 |
| DeepSeek V4 Flash 284B · 13B active | doesn't fit at these settings | 130.8 t/sQ2_K |
| Qwen3.6 27B | 7.9 t/sFP16 | 18.3 t/sFP16 |
| Gemma 4 26B · 4B active | 19.0 t/sFP16 | 49.3 t/sFP16 |
Change the context, RAM or model in the calculator: M3 Ultra 96GB · NVIDIA RTX Pro 6000 Blackwell
Every report on each card, with its source: M3 Ultra 96GB · NVIDIA RTX Pro 6000 Blackwell