self-hosted/ai
§01·model · /models

Qwen3.5 27B

multimodalactiveApache-2.0

27B dense LLM by Alibaba (Qwen3.5). Apache-2.0.

§02·GPUs that run this model
2 total
GPUVRAMSeriesBest speedMin VRAMWorksBenchmarksRecipe
RTX 509032GB5058.8tokens/s2recipecheck ↗
RTX 309024GB3033.5tokens/s24GB2recipecheck ↗

benchmarked·~ runs via recipe (not benchmarked)· untested·doesn't fit