M3 Ultra 256GB vs NVIDIA RTX 5090 for local LLMs
15 reports on the M3 Ultra 256GB and 170 on the NVIDIA RTX 5090
Which is faster for local LLMs?
On Qwen3.8 27B at 4-bit, the M3 Ultra 256GB has 14.0 generation tokens/s (one run); prompt processing 93.1 tokens/s (one run); the NVIDIA RTX 5090 has 80.0 generation tokens/s (7 runs). 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 2.2x higher on the NVIDIA RTX 5090 (819 GB/s on the M3 Ultra 256GB, 1,792 on the NVIDIA RTX 5090).
- Memory
- 256GB unified
- Memory bandwidth
- 819 GB/s
- FP16 compute
- 28.4 TFLOPS
- Reports
- 15
- Memory
- 32GB
- Memory bandwidth
- 1,792 GB/s
- FP16 compute
- 419 TFLOPS
- Reports
- 170
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 256GB | NVIDIA RTX 5090 |
|---|---|---|
| Qwen3.8 27B4-bit | 14.0one runPP 93.1 (one run)Ollamacontext not stated | 80.07 runsllama.cpp, TensorSharp +1, 2 not stated40K to 256K context, 2 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 | M3 Ultra 256GB | NVIDIA RTX 5090 |
|---|---|---|
| Qwen3.8 27B | 8.0 t/sFP16 | 44.9 t/sQ6_K |
| Qwen3.8 125B · 6B active | 27.3 t/sQ8_0 | 20.3 t/sQ2_K, experts in RAM |
| Qwen3.6 35B · 3B active | 28.7 t/sFP16 | 163.0 t/sQ5_K_M |
| DeepSeek V4 Flash 284B · 13B active | 36.0 t/sQ4_K_M | doesn't fit at these settings |
| Qwen3.6 27B | 8.0 t/sFP16 | 44.9 t/sQ6_K |
| Gemma 4 26B · 4B active | 21.5 t/sFP16 | 114.5 t/sQ6_K |
| DeepSeek V4.1 Flash 552B · 16B active | 48.0 t/sQ2_K | doesn't fit at these settings |
Change the context, RAM or model in the calculator: M3 Ultra 256GB · NVIDIA RTX 5090
Every report on each card, with its source: M3 Ultra 256GB · NVIDIA RTX 5090