ComfyUI update doubles Qwen LLM token rates on consumer GPUs, up to +251% on RTX 5090
Lexius2129 · reddit · 2026-11-12
The ComfyUI team shipped optimizations that dramatically speed up local text generation with Qwen models — most local users still run 12GB RTX 3060 cards, so the team keeps optimizing for consumer hardware.
Benchmark data from the official blog (v0.40.0 vs v0.44, Qwen 3.x W4A8):
- RTX 5090: Qwen3.5 9B story generation 106 → 219 tokens/s (+106%); Qwen3.8 27B story +91%, image captioning +120%, and 12k-context runs 40 → 142 tokens/s (+251%).
- RTX 5080: gains of 41–136%.
Text generation matters in creative workflows for prompt building and local harnesses like Comfy Agent.
Related event: ComfyUI Boosts Local Qwen Text Generation by Up to 251%(2 posts)→
More from Infra
- SGLang Q4 roadmap: 10+ accelerator families supported, plugin system makes new backends first-class — ying11231 · 2026-11-12
- SGLang ships Unified Memory: one GPU memory arena for every token lifetime, no partitioning knob — ying11231 · 2026-11-12
- Polymarket odds hit 72% for Jensen Huang's removal by tomorrow amid board investigation — Polymarket · 2026-11-12
- Cerebras explains wafer-scale yield: independent voltage tuning across 84 die regions plus redundant cores — Sethwinterroth · 2026-11-12
- gVisor joins CNCF with Modal as maintainer, running 40M+ containers daily — charles_irl · 2026-11-12
- Hugging Face Launches Open-Source Llama for Windows, Bringing No-Setup Local AI to the Masses — unixterminal · 2026-11-12