Qwen
Qwen
Qwen/Qwen3.5-9B

Runs

View all benchmark runs for this model family.

	Quant
M4 Max	q8_0	llama.cppb8680	35.6 tok/s	698.2 tok/s	14.60 GB 41%
M4 Max	OptiQ-4bit	mlx_lm0.31.2	60.9 tok/s	707.4 tok/s	8.11 GB 23%
M4 Max	q8_0	llama.cppb8680	35.9 tok/s	696.6 tok/s	9.20 GB 26%
M4 Max	bf16	llama.cppb8680	21.3 tok/s	707.1 tok/s	17.40 GB 48%
M4 Max	4bit	mlx_lm0.31.1	66.5 tok/s	633.1 tok/s	7.32 GB 20%
M4 Max	bf16	mlx_lm0.31.1	21.9 tok/s	754.4 tok/s	19.00 GB 53%
M2 Pro	OptiQ-4bit	mlx_lm0.31.1	28.6 tok/s	192.8 tok/s	8.16 GB 26%
M2 Pro	OptiQ-4bit	mlx_lm0.31.1	30.1 tok/s	201.0 tok/s	8.16 GB 26%
M2 Pro	OptiQ-4bit	mlx_lm0.31.1	27.9 tok/s	184.0 tok/s	8.18 GB 26%
M2 Pro	OptiQ-4bit	mlx_lm0.31.1	30.0 tok/s	204.2 tok/s	8.06 GB 25%