ggml @ggml_org · 5d, Nov 4, 2026, 4:16 PM llama.cpp v0.9.0github.com/ggml-org/llama… github.com Release v0.9.0 · ggml-org/llama.cpp Overview This release adds OCP FP8 E4M3 quantization support, a new flash_attn_ext_rows op that makes the unified KV cache efficient for multi-sequence inference, and a generic batched tensor copy ...
ggml @ggml_org · 5d, Nov 4, 2026, 12:55 PM ggml v0.29.0github.com/ggml-org/ggml/… github.com Release v0.29.0 · ggml-org/ggml Overview ggml v0.29.0 adds an OCP FP8 E4M3 tensor type, a new ggml_flash_attn_ext_rows op where each query slice attends to its own list of K/V rows, a batched tensor copy API, and initial support ...
ggml @ggml_org · Oct 28, Oct 28, 2026, 3:09 PM llama.cpp v0.8.0github.com/ggml-org/llama… github.com Release v0.8.0 · ggml-org/llama.cpp Overview This release adds the LiquidAI LFM3 family (text, vision, audio and omni) together with a new experimental streaming-audio API in mtmd, and wires realtime audio in and out of the server th...
ggml @ggml_org · Oct 28, Oct 28, 2026, 11:38 AM ggml v0.28.0github.com/ggml-org/ggml/… github.com Release v0.28.0 · ggml-org/ggml Overview This release adds a fused sparse-attention masking operation (GGML_OP_TOP_K_MASK) with CPU, CUDA and Metal implementations, plus accessors to read the precision configured on a compute nod...
ggml @ggml_org · Oct 15, Oct 15, 2026, 5:22 PM llama.cpp v0.7.0github.com/ggml-org/llama… github.com Release v0.7.0 · ggml-org/llama.cpp Overview llama.cpp v0.7.0 adds the LiliCorr speculative decoding technique and MTP for GLM5-Next, a new GPU cache for MoE experts kept in host memory (--moe-cache-mib, up to ~2.2x decode speedup wh...
