Vaya's read on this project
Problem, audience, market, and the verdict — sign in to see it.
Updates
June 2026
- jinja, chat: add --reasoning-preserve flag (#25105)
- Revert "ui: fix accessibility for hover-gated interactive elements as…
- ui: fix stop and reasoning skip in single-model mode (#25084)
- dflash: refactor draft model conversion (#25110)
- chat : implement minicpm5 parser (#24889)
- jinja: add --dump-prog for debugging (#25086)
- spec : add DFlash support (#22105)
- common : allow --offline in llama download (#25091)
- logs : reduce v2 (#25078)
- opencl: flash attention improvement (#25069)
- [CUDA] Added a cudaMemcpy2DAsync fast path to ggml_cuda_cpy (#25057)
- sycl : fix failed ut cases of norm (#25044)
- vulkan: fix step operator for 0 input (#25036)
- binaries : Improve rpc-server and export-graph-ops names. (#25045)
- ci : add windows-openvino to check-release (#25022)
- tests : fix test-chat-template --no-common option (#25075)
- app : allow --version, --licenses & --help (#25054)
- sched : reintroduce less synchronizations during split compute (#20793)
- devops : add llama in all docker images (#25035)
- arg: fix handling --spec-draft-hf and --hf-repo-v (#25043)
October 2025
- Release —b6795
- Release —b6794
- Release —b6792
- Release —b6791
- Release —b6788
- Release —b6783
- Release —b6782
- Release —b6781
- Release —b6779
- Release —b6776