Qwen
Qwen
Qwen/Qwen3.5-35B-A3B

Runs

View all benchmark runs for this model family.

	Quant						Actions

	Quant		Actions

	Quant						Actions
M5 Max	4bit	mlx_lm0.31.2	127.1 tok/s	4,044.5 tok/s	21.00 GB 33%
M5 Max	8bit	mlx_lm0.31.2	101.5 tok/s	3,939.1 tok/s	37.00 GB 29%
M2 Max	4bit	mlx_lm0.31.2	77.9 tok/s	721.8 tok/s	21.00 GB 33%
M2 Max	4bit	mlx_lm0.31.2	75.5 tok/s	664.2 tok/s	21.00 GB 33%
M5 Max	4bit	mlx_lm0.31.2	138.5 tok/s	4,334.9 tok/s	21.00 GB 16%
M4 Pro	4bit	mlx_lm0.31.2	72.8 tok/s	820.6 tok/s	21.00 GB 44%
M4 Max	4bit	mlx_lm0.31.2	104.0 tok/s	1,461.7 tok/s	21.00 GB 58%
M4 Max	q4_k_m	llama.cppb8680	60.1 tok/s	1,099.2 tok/s	20.65 GB 57%
M2 Max	4bit	mlx_lm0.31.1	62.5 tok/s	509.3 tok/s	21.00 GB 22%
M4 Max	4bit	mlx_lm0.31.1	105.1 tok/s	1,412.7 tok/s	21.00 GB 58%

	Quant		Actions
M5 Max	4bit	127.1 tok/s
M5 Max	8bit	101.5 tok/s
M2 Max	4bit	77.9 tok/s
M2 Max	4bit	75.5 tok/s
M5 Max	4bit	138.5 tok/s
M4 Pro	4bit	72.8 tok/s
M4 Max	4bit	104.0 tok/s
M4 Max	q4_k_m	60.1 tok/s
M2 Max	4bit	62.5 tok/s
M4 Max	4bit	105.1 tok/s