community pool / model answer

Qwen3.6-35B-A3B-4bit

chat · mlx-community/Qwen3.6-35B-A3B-4bit

106.4 tok/s · model median · 2 runs in the model record.

Reference rigQuantizationMedian resultBasisTTFT (ms)Peak VRAMMax contextEngines
M5 Pro 64GB4-bit103.9 tok/sreported22919.9 GB2,048mlx
Max 128GB4-bit109 tok/sreported-- GB2,048mlx

Context tested

The table above is the tested context for each exact rig and quantization cell. Empty fields remain empty rather than being filled with an estimate.