Strata Runs 100 tks in Agentic Workloads but Shows Looping and Port-Conflict Issues
GodComplecs · reddit · 2026-10-14
A Reddit user shares hands-on findings on the Strata model: it hits 100 tokens/s in agentic workloads versus 40-60 tks for similar-size Qwen 27B on llama.cpp, but exhibits odd looping issues in chat and repeatedly tries to use ports already in use, which Qwen 27B Q4 UD did not. Community feedback on real downsides remains split between 'no loss' and 'huge loss.'
More from Infra
- Dev complains HDDs grind his server to a halt as NVMe markup hits 67,000% — haydendevs · 2026-10-14
- Sniffed the network traffic of Ollama and LM Studio: chats stay local, apps don't — TheOyinbooke · 2026-10-14
- First real-world look at Google's Ironwood TPU rack at Google Cloud event — firstadopter · 2026-10-14
- Samsung shows off HBM5 samples and Custom HBM with xPU stacked design — firstadopter · 2026-10-14
- AWS launches Lambda Web Functions with up to 128 concurrent requests per environment — tlakomy · 2026-10-14
- Bosch and Rittal show massive manifolds for MW liquid cooling and 800VDC — BenBajarin · 2026-10-14