community pool / model answer

Qwen3.6-35B-A3B-mxfp4

chat · mlx-community/Qwen3.6-35B-A3B-mxfp4

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

Reference rigQuantizationMedian resultBasisTTFT (ms)Peak VRAMMax contextEngines
M5 Max 128GB4-bit143.3 tok/sreported-18.9 GB2,048mlx
M5 Pro 64GB4-bit106.1 tok/sreported22518.9 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.