Back to home
Consumer GPU
2 articles tagged with this topic
QwenvLLM
Qwen 27B Hits 218 Tokens/sec on Two RTX 3090s — Local AI Cost Curve Drops
Developer hits 218 tokens/sec running Alibaba's 27B Qwen on two consumer RTX 3090s. H100-cluster workloads now run on consumer GPUs. Local AI hardware
Aug 192 min read
QwenRTX 3090
Qwen 27B Hits 99 tps on a $200 RTX 3090 — Personal Local AI Arrives
GitHub user syv-ai got Alibaba's Qwen 27B running at 99 tokens/sec on a used RTX 3090 (24GB VRAM, ~$200) — consumer GPUs can now smoothly handle mid-s
Aug 182 min read