← Back to live feed1 story
Llama.cpp v0.6.0 introduces multimodal support for decision models, including the Clef and Clef-Flash models. The update also includes performance enhancements for the Metal API to accelerate on-device execution on Apple hardware.
Content history (2)
- 2026-10-08 · Summary · vi · Replaced stored textLlama.cpp vừa phát hành phiên bản v0.6.0 với tính năng hỗ trợ decision model đa phương th…Các mô hình Clef và Clef-Flash hiện đã có sẵn để triển khai cục bộ trên phiên bản mới nhấ…
- 2026-10-08 · Title · vi · Replaced stored textLlama.cpp hỗ trợ multimodal decision model trong phiên bản v0.6.0Llama.cpp hỗ trợ multimodal decision model trong v0.6.0
Sign in to suggest edits
Key sources
- SOURCE@ngxson“already 5 open models supported, from 144M to 27B”x.com
- SOURCE@ggerganov“The /v1/systemone endpoint is available in the latest llama builds”x.com
- SUPPORT@mervenoyann“new Llama.cpp release ships with (multimodal!) Jev-like models support, performance upgrade for Metal and more!”x.com
- SUPPORT@vanstriendaniel“Decision models now run on device in llama.cpp. Free, fast, private!”x.com
- SUPPORT@victormustar“tried Laya (421M) on llama.cpp in a little apple-picking benchmark, it's almost as good as Qwen 4B at 12x the speed”x.com
- SUPPORT@lucataco“You can now run Clef via llama.cpp!”x.com
- SUPPORT@calebfahlgren“run the popular open jev alternatives via llama.cpp! it's also compatible with @typesafeai sdk”x.com
- SUPPORT@fffiloni“Give Julia-1 a situation: it ranks options, estimates a yes/no probability, and scores an ordered scale. No generated text.”x.com