Headline
- Support Qwen3-Next GGUFs by upgrading the llama.cpp backend version.
- Overhaul the model download manager to support resuming partial downloads and provide more user feedback.
What's Changed
- Documentation cleanup by @jeremyfowers in #618
- Replace conda with venv in CI and docs by @jeremyfowers in #632
- New download manager and support for Qwen3-Next by @jeremyfowers in #625
- fix: Serve favicon from root by @VladimirVLF in #622
Full Changelog: v9.0.4...v9.0.5