GLM-5.3-Flash quantized to 2.0bpw runs on a single DGX Spark at 18-25 tok/s with vision
pbaylies · x · 2026-09-08
Developer 0xSero released an EXL3 quantization (TR3, 2.0bpw, no pruning) of GLM-5.3-Flash sized for a single NVIDIA DGX Spark. It's validated working at 18-25 tok/s with vision enabled, weights are up on Hugging Face with full chat template config, and the author is soliciting community testing feedback.
More from Infra
- Anthropic S-1: training infra edge cut per-task cost 40% and doubled revenue growth — BenBajarin · 2026-10-26
- Rewriting scaling law study: optimal synthetic data share rises from 10% to 32% with compute — heghbalz · 2026-10-26
- US states halt new data centers as 72% of Americans oppose local builds — YvesMulkers · 2026-10-26
- $5M Seed Round Can't Afford One DeepSeek Training Run: GPU-less AI Startup's Three Paths — FinanceYF5 · 2026-10-26
- HBM demand shifts: low-stack inference plans shelved, Rubin to go 12-high only — BenBajarin · 2026-10-26
- CircuitHub unveils The Grid, a factory-scale robot making PCBs in the US twice as fast as Shenzhen — garrytan · 2026-10-26