M5 Ultra 256GB vs AMD Strix Halo 128GB for local LLMs
30 reports on the M5 Ultra 256GB and 320 on the AMD Strix Halo 128GB
Which is faster for local LLMs?
On Qwen3.8 27B at 4-bit, the M5 Ultra 256GB has 50.0 generation tokens/s (one run); prompt processing 1800.0 tokens/s (one run); the AMD Strix Halo 128GB has 13.1 generation tokens/s (4 runs); prompt processing 378.2 tokens/s (3 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 4.8x higher on the M5 Ultra 256GB (1,229 GB/s on the M5 Ultra 256GB, 256 on the AMD Strix Halo 128GB). Across 2 shared models with generation measurements on both cards, the M5 Ultra 256GB's median is higher on 2 and the AMD Strix Halo 128GB's on 0.
- Memory
- 256GB unified
- Memory bandwidth
- 1,229 GB/s
- FP16 compute
- not on record
- Reports
- 30
- Memory
- 128GB unified
- Memory bandwidth
- 256 GB/s
- FP16 compute
- 59 TFLOPS
- Reports
- 320
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 | M5 Ultra 256GB | AMD Strix Halo 128GB |
|---|---|---|
| Qwen3.8 27B4-bit | 50.0one runPP 1800.0 (one run)oMLX8K context | 13.14 runsPP 378.2 (3 runs)llama.cpp, 1 not stated32K context, 3 not stated |
| Qwen3.5 122B · 10B active4-bit | 74.7one runPP 2710.5 (one run)MLXcontext not stated | 22.4one runPP 287.0 (one run)engine 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 | M5 Ultra 256GB | AMD Strix Halo 128GB |
|---|---|---|
| Qwen3.8 27B | 14.9 t/sFP16 | 3.5 t/sFP16 |
| Qwen3.5 122B · 10B active | 31.0 t/sQ8_0 | 7.7 t/sQ6_K |
| Qwen3.8 125B · 6B active | 50.8 t/sQ8_0 | 12.6 t/sQ6_K |
| Qwen3.6 35B · 3B active | 53.5 t/sFP16 | 12.6 t/sFP16 |
| DeepSeek V4 Flash 284B · 13B active | 66.9 t/sQ4_K_M | 33.5 t/sQ2_K |
| Qwen3.6 27B | 14.9 t/sFP16 | 3.5 t/sFP16 |
| Gemma 4 26B · 4B active | 40.1 t/sFP16 | 9.5 t/sFP16 |
| DeepSeek V4.1 Flash 552B · 16B active | 89.4 t/sQ2_K | doesn't fit at these settings |
Change the context, RAM or model in the calculator: M5 Ultra 256GB · AMD Strix Halo 128GB
Every report on each card, with its source: M5 Ultra 256GB · AMD Strix Halo 128GB