Qwen-Image-2.1 gets GGUF quantization, could run text-to-image on a Snapdragon 865 phone
ResidentAping · reddit · 2026-09-25
A Reddit user notes that both Qwen-Image-2.1 and Qwen3-VL-8B now ship in GGUF format, speculating that a Q4KM quantization could run text-to-image locally on Android devices like a Snapdragon 865 — similar to how Z-Image-Turbo does. The post includes several sample images generated with prompts from Civitai, with the author admitting it would be very slow and calling it "just an idea."
More from Infra
- Altman says OpenAI aims to replace Nvidia in the medium to long term — zephyr_z9 · 2026-11-11
- Chinese University Bought Nvidia H200 Chips Despite US Blacklist — pstAsiatech · 2026-11-11
- Baidu's ERNIE Evolve tops public bests on NanoGPT Speedrun, brings RSI to real inference optimization — xiaohu · 2026-11-11
- Baidu says RSI auto-optimized ERNIE inference, 6.24x decode speedup on Kunxin P800 — xiaohu · 2026-11-11
- Nvidia's supply commitments jump from $119B to $279B in one quarter, locking up 2027-28 memory — tengyanAI · 2026-11-11
- Train 27B models on a 24GB GPU: ternary optimizer cuts memory to 1 byte/weight — oguzeldereli · 2026-11-11