LM Studio on the M6 32GB
On the M6 32GB, LM Studio runs Qwen3.6 35B · 3B active at 4-bit at a median 63.8 tokens per second for generation, from 1 plain run (32K context).
Speed by model on the M6 32GB
Plain runs only: one device, one request, no speculative decoding, the whole model in memory. Each row is one model size at one bit class, with the other engines' plain runs of the same model on this card beside it. Context length and build still differ between runs, and each side shows its own.
| Model | LM Studio | Other engines |
|---|---|---|
| Qwen3.6 35B · 3B active 4-bit | 63.8 t/sone report, 32K context | no other engine reported |
15 LM Studio reports on the M6 32GB
Newest first. A run on several cards, with speculative decoding, with part of the model in system RAM or with batched requests says so, since each describes a different setup.
- Qwen3.5 9B30.3 t/s
- Gemma 4 31B7.5 t/s
- Gemma 4 12B18.8 t/s
- Gemma 4 26B · 4B active48.2 t/s
- gpt-oss 21B · 3.6B active45.2 t/s
- Muse8.6 t/s
- Nemotron 345.0 t/s
- Ornith1.5 35B · 3B active61.5 t/s
- Qwen3.8 27B20.2 t/s
- Qwen3.8 27B19.5 t/s
- Qwen3.8 27B · Q4_K_M11.8 t/s
- Ornith1.5 35B · 3B active · Q4_K_M43.7 t/s
- Nemotron 3.5 Lightning 30B · 3B active · UD-Q3_K_XL42.3 t/s
- Nemotron 3.5 Lightning 30B · 3B active · Q4_K_M45.1 t/s
- Qwen3.6 35B · 3B active · MLX 4-bit63.8 t/s
Frequently asked
How fast is LM Studio on the M6 32GB?
On the M6 32GB, LM Studio runs Qwen3.6 35B · 3B active at 4-bit at a median 63.8 tokens per second for generation, from 1 plain run (32K context).