GLM-5.3 320B (18B active) Flash
on M3 Ultra 256GB · 300,000 ctx
Use cases
agenticcodingtool-use
Summary
User reports GLM-5.3-Flash Q4 at 37.4 t/s end to end at 300k context on an M3 Ultra, against 21.6 t/s for stock ds4.
Setup is a ds4 branch tuned for the M3 Ultra, decoding serially in a single stream without the dflash drafter, with prefill at 550 t/s at 62k.
Serial decode rises from 29 to 40 t/s at short context and from 24 to 38 at 62k. In a claude code harness at ~200k depth it averages over 38 t/s output and reaches 60 t/s writing SQL, where the optional dflash drafter adds +20 to +50 percent.