Search: “models”

spec : auto-detect mtp draft model type (#27005) Website: https://llama.app macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFramew

mtmd: support pocket-tts (#26871) adapt the api text model ok working impl, need verify and clean up mtmd: build the pocket-tts transposed convolutions as GEMM + col2im ggml_conv_transpose_1d has no g

chat : tighten bare function parsing for Qwen models (#26793) Website: https://llama.app macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64)

common: Add CLI > ENV > models-presets > INI precedence CLI flags have the highest precedence ENV vars have the second-highest precedence System and User configs have the lowest precedence Linux/BSD/M

server: (router) do not evict busy models (#26567) Website: https://llama.app macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFram

tts: account for the vocoder pass in the timings line (#26733) get_output runs the waveform work the pipeline defers to it, from a single trailing window to a full pass depending on the model. Measuri

model-loader : fix quantized reshaped tensor strides (#26672) Website: https://llama.app macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64)

model : support MTP in GLM-4.7-Flash (#24868) Website: https://llama.app macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFramework

model : allow reshape of tensors during load (#26531) Website: https://llama.app macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCF

models : fix dflash wo_a reshape on load (#26577) Website: https://llama.app macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFrame

model: MTP support for Qwen3-Next (#25589) mtp for qwen3nex fix for python type-check Fix to compute num_mtp from directly mtp layer define opt_num_mtp_layers in _QwenMtpMixin and fix some comments Fi

CUDA: Add backend sampler for penalties sampler (#25262) sampling: enhance penalty handling in common_sampler_init Set default value for penalty_last_n based on model context if not specified. Ensure

model: M3: Move MSA into a new memory implementation (#26338) Move MSA logic from llama-kv-cache into llama-kv-cache-msa cont : minor cont : ws fix Co-authored-by: Georgi Gerganov ggerganov@gmail.com

model : load MiMo V2 MTP tensors only if used (#26412) Website: https://llama.app macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XC

common: support the DSpark sidecar resolution (#26458) The dspark- files resolve like the other speculative sidecars: the -hfd tag applies to them, a requested sidecar resolves without a full model at

llama : load MTP tensors only if they are really used (#26296) llama : load MTP tensors only if they are really used llama : skip loading MTP (if not used) in remaining models that support MTP Co-auth

spec: add eagle3-v3 support for gpt-oss model (#25794) Website: https://llama.app macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XC

model: Add Laguna-S-2.1 LLM_TYPE (#26233) Website: https://llama.app macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCFramework Lin

mtmd: support MiMo-V2.5 audio input (RVQ-based model) (#26190) gguf converter for mimo audio fix conv cpp impl nits nits 2 Website: https://llama.app macOS/iOS: macOS Apple Silicon (arm64) macOS Apple

What's Changed mlx update by @dhiltgen in #17332 model/parsers: finalize incomplete GLM tool calls by @dhiltgen in #17250 docs: update retirements by @mxyng in #17289 model: align Laguna with upstream

vulkan: support type-aligned GET_ROWS (#28253) vulkan: fall back to CPU for GET_ROWS with misaligned offsets The Vulkan GET_ROWS shader asserts when a tensor's backing-buffer offset plus view_offs is

mtmd : add const in various places (#28307) mtmd : mark context as const in more methods Mark mtmd_context as const in: mtmd_bitmap_init_lazy mtmd_tokenize mtmd_tokenize_from_parts mtmd_helper_support

qwen4exp: fix seq_cp, block position keying, mtmd input, cuda abort, add tests (#27941) qwen4exp: follow up fixes -kvu NaN collapse fix Assisted-by: Claude indexer cache ext.x/ext.y restore fix Assist

opencl: use a better matmul path on two Adreno GPU generations (#27640) opencl: default the Adreno xmem F16xF32 GEMM on for X2E kernel_mul_mm_f16_f32_l4_lm is the slowest matmul this backend has on Ad

What's Changed launch: add Muse Code integration by @dhiltgen in #17594 model/renderers: match Muse Glimmer reasoning template by @dhiltgen in #17732 launch: add DeepSeek Harness integration by @Parth

SYCL: add oneMKL GEMM flash attention for XMX-accelerated prompt proc… (#25025) SYCL: add oneMKL GEMM flash attention for XMX-accelerated prompt processing fattn-mkl: fix interleaved dst layout in nor

common: infer the speculative type from the draft repo sidecars (#25989) With -hfd pointing to a repo that ships mtp-/dflash-/eagle3- sidecars and no --spec-type given, the draft resolved to a full mo

This refactors the existing integration tests into 3 priumary groups: fast, release, and library. It also refines some of the release tests to drop some of the older models and pick up newer models, w

server : refactor prompt cache state ownership (#25649) server : clear checkpoints upon prompt clear server : move the prompt state data to the server_prompt_cache Assisted-by: pi:llama.cpp/Qwen3.6-27

The b9761 release of llama.cpp includes server improvements with model downloading moved to a dedicated process and real-time model load progress tracking via /models/sse endpoint. The release feature

B9544 is an intermediate build release of llama.cpp, an open-source C/C++ library for running large language model inference. Llama.cpp uses the GGUF model format and supports multiple hardware backen

Ollama 0.30 provides improved compatibility and performance using llama.cpp, augments the MLX engine on Apple Silicon for broader hardware support, and brings support for a wider range of models inclu

Ollama v0.30.1 addresses an issue where the LLM component now ignores Server-Sent Events (SSE) ping comments from llama-server, resolving problem #16443. This fix improves the stability and reliabilit

b9391 is a build release of llama.cpp, an open-source C/C++ inference framework for running large language models locally. llama.cpp provides lightweight, optimized model inference with support for mu

Release b9142 is an intermediate build of llama.cpp, a C/C++ library for running large language model inference on consumer hardware. llama.cpp performs inference on various large language models and

Ollama v0.20.4-rc2 is a release candidate that addresses a compatibility issue with Flash Attention (FA) for the Gemma 4 model on older GPUs. CUDA versions older than 7.5 lack the support needed t...

llama : do not map foreign ranges into a backend buffer A buffer created from a host pointer pins everything that it wraps. When several contexts share the same file (lazy tensors, MoE tensors kept on

metal : fix MUL_MAT+ADD fusion when the residual is itself a MUL_MAT (#30100) metal : fix MUL_MAT+ADD fusion when the residual is itself a MUL_MAT ggml_metal_op_mul_mat_mma picks the residual of a fus

imatrix: calculate activation-based statistics for new format (GGUF) imatrices (#14891) Use activations to calculate the stats Determine calculation mode Compute entropy for activations Compute cosine

llama: llama_prefetch_rows (#29599) llama: llama_prefetch_rows llama: support row prefetch on Windows Apply the Windows port contributed by @praneshgo unchanged. Source: #29599 (comment) avoid exposin

cli: exit on stdin EOF and drop the console wide Ctrl+C broadcast (#29722) cli: exit on stdin EOF and drop the console wide Ctrl+C broadcast On Windows the simple input reader sends CTRL_C_EVENT to ev

What's Changed feat: allow ten web searches per response by @ParthSareen in #18602 MLX: version bump by @dhiltgen in #18651 llama.cpp: version bump b11232 by @dhiltgen in #18652 create: support explic

context : do not re-reserve the scheduler when toggling causal_attn (#28751) context : do not re-reserve the scheduler when toggling causal_attn llama_context::set_causal_attn() marks the scheduler to

server : support typed content (vision/audio/video) input for /v1/embeddings endpoint (#29556) server : support multimodal input for /v1/embeddings (Qwen3-VL-Embedding) Accept the OpenAI-style wrapped

Add manifest-list storage so runner-specific manifests can coexist under one tag while preserving existing v1 tags as best-effort downgrade anchors. Show/list/copy/remove/pull/push now understand runn

hexagon: overhaul of buffer and DMA handling to support 64bit mappings + improvements (#29197) hex-dma64: enable support extended buffer mappings and 64bit dma hex-dma64: expand binary ops to support