M5 24GB vs M6 32GB for local LLMs
2 reports on the M5 24GB and 22 on the M6 32GB
Which is faster for local LLMs?
Nobody has reported the same model at the same quant on both cards yet, so there's no measured head-to-head. Memory bandwidth, which caps how fast a card can write, is 11% higher on the M6 32GB (154 GB/s on the M5 24GB, 171 on the M6 32GB). The estimates below compare them on the models people run most.
M5 24GB
- Memory
- 24GB unified
- Memory bandwidth
- 154 GB/s
- FP16 compute
- not on record
- Reports
- 2
M6 32GB
- Memory
- 32GB unified
- Memory bandwidth
- 171 GB/s
- FP16 compute
- not on record
- Reports
- 22
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.
| Model | M5 24GB | M6 32GB |
|---|---|---|
| Qwen3.8 27B | 14.7 t/sQ2_K | 10.5 t/sQ4_K_S |
| Qwen3.8 125B · 6B active | doesn't fit at these settings | doesn't fit at these settings |
| Qwen3.6 35B · 3B active | 48.7 t/sQ2_K | 38.2 t/sQ3_K_M |
| DeepSeek V4 Flash 284B · 13B active | doesn't fit at these settings | doesn't fit at these settings |
| Qwen3.6 27B | 14.7 t/sQ2_K | 10.5 t/sQ4_K_S |
| Gemma 4 26B · 4B active | 32.8 t/sQ3_K_S | 23.2 t/sQ4_K_M |
Change the context, RAM or model in the calculator: M5 24GB · M6 32GB
Every report on each card, with its source: M5 24GB · M6 32GB