Qwen
Qwen/Qwen3-Coder-30B-A3B-Instruct

OverviewRuns

Quantizations

QuantQuantized bySizeDecodePrefillScoreActions
4bitMLX's logo.
LM Studio
LM Studio		16.0 GBN/AN/AN/A
4bitMLX's logo.
MLX Community
MLX Community		16.0 GB22.8 tok/s147.8 tok/sRuns poorly

Device Comparison

Results include trials with 4,096 input tokens and 1,024 output tokens only.

Decode / Prefill Speeds

2 devices