Report
llama.cpp v0.6.0 adds Clef support for text and vision
The release announcement touts Qwen3.8-Flash-Next support, major Metal performance gains and a new `llama_batch_ext` API.
TLDR
The v0.6.0 release announcement lists Clef text and vision support, Qwen3.8-Flash-Next support, what it calls a massive Metal performance improvement, a new llama_batch_ext API and a refreshed website. Another user says Clef can now run via llama.cpp and shares an Ollama-versus-llama.cpp speed test on an M5 Max MacBook Pro.
Combined views
2.1K
3 Sources, first seen ago
