chore: sync with upstream ggml-org/llama.cpp (106 commits)#46

Merged

TheTom merged 89 commits into

feature/turboquant-kv-cachefrom

pr/upstream-sync-april

Apr 2, 2026

Commits on Mar 27, 2026

CANN: update docker images to 8.5.0 and improve CANN.md (#20801)
KokerZhou
authored
metal : Fix dimension constraint violation in matmul2d descriptor (#21048)
lathrys-at
authored
Send reasoning content back to the model across turns via the reasoning_content API field (#21036)
ServeurpersoCom
authored
completion : Fix segfault on model load failure (#21049)
mtmcp
authored
completion : session_tokens insert range in completion tool (no-op → correct) (#20917)
mtmcp
authored
rpc : proper handling of data pointers to CPU buffers (#21030)
rgerganov
authored
server: add built-in tools backend support (#20898)
Xuan-Son Nguyen
authored
mtmd: add more sanity checks (#21047)
Xuan-Son Nguyen
authored
server: remove the verbose_prompt parameter (#21059)
aisk
authored
webui: Improve Chat Messages initial scroll + auto-scroll logic + add lazy loading with transitions to content blocks (#20999)
allozaur
authored
hexagon: support for IQ4_NL and MXFP4 (#21018)
njsyw1997
authored
server: Introduce LLAMA_BUILD_WEBUI build flag to allow disabling the embedded web ui (#20158)
kushagharahi
authored
common : inhibit lazy grammar sampler while reasoning is active (#20970)

aldehir
and
pwilkin
authored

Commits on Mar 28, 2026

Commits on Mar 30, 2026

Commits on Mar 31, 2026

Commits on Apr 1, 2026

Commits on Apr 2, 2026