Search: “models”

model-saver : fix expert shared/chunk FFN length key clobber (#26693) The saver called add_kv with LLM_KV_EXPERT_SHARED_FEED_FORWARD_LENGTH twice, the second time passing n_ff_chexp. gguf_set_val_u32

feat: add GLM5Next MTP, optimize (#29928) llama : add GLM5-Next NextN (MTP) graph Build the GLM5-Next multi-token-prediction head as graph_mtp: the NextN block embeds enorm(tok)+hnorm(h) through eh_pr

ggml-openvino: skip unselected graph branches and support DUP Upstream #29622 adds a mixed token/embd branch to every input embedding graph through ggml_build_forward_select(). Its nodes are not flagg

ggml: refactor selective expert copying to user code (#29943) ggml: refactor selective expert copying to user code tests: enroll two models into selective expert copy test tests: use deepseek2 as test

ci : fix Models Backend Check by shortening the hrm_text fixture (#29744) The fixture recycles its two blocks over 8 cache slots, so the fp16 error builds up past the 1e-4 NMSE bound on the Vulkan T4

What's Changed app: fix ChatGPT model selector spacing mlxrunner: Evict prefix cache snapshots from the active conversation mlxrunner: check system free memory and wait for evicted runners before load

model: add NVIDIA Nemotron-3-Puzzle-75B-A9B (NemotronHPuzzle) support (#25444) hparams: add per-layer n_ff_exp/n_expert_used arrays with scalar-or-array loading G1/G2 infrastructure for variable-per-l

AVX2: Speed up large batch size prompt processing of IQ models (#27402) Batched gemm for grid IQ quants Style updates and a bit more performance Clean up comments Move code around Vectorize IQ panel d

Qwen 3.8 - 27B model support This release adds the support of Qwen 3.8 27B. For Apple Silicon devices, Ollama has in particular optimized for maximum performance and output quality suitable for repeat

What's Changed Qwen3.5 is faster on Apple GPUs: the MLX engine now uses the model's MTP head for speculative decoding automatically /v1/chat/completions streaming now matches OpenAI's wire format: rol

What's Changed New interactive agent experience: running ollama now launches an agent to help you code and delegate work ❯ ollama Ollama 0.32.0 ▸ Chat, Code, & Work (glm-5.2:cloud) Chat with models, c

model : support for Prism Bonsai 2 27B (#29600) Runtime support for Prism Bonsai 2 27B Assisted-by: Claude Code address prism hadamard runtime feedback move hadamard tensors into method, define folded

What's Changed ollama launch claude no longer shows a warning about Claude connectors by @hoyyeva in #18878 ollama launch codex no longer warns about service tiers by @hoyyeva in #18879 Paused backgro

ci: fix Models Backend webgpu by supporting GGML_OP_DUP (#30216) The mixed batch path of PR-29622 writes the token rows with set_rows into a dup of the embeddings. WebGPU did not support DUP, so the d

mtmd: add cohere2 vision support (#30062) cohere2 vision model address comments remove redundant mapping follow existing patterns fused linear_1 Website: https://llama.app Attestations: https://github

kv-cache: fix restoring mismatched KV cache rotation by saving exact rotation metadata (#28498) kv-cache: save exact KV rotation metadata, reject restoring mismatched rotation tests : move the state r

common : prepare load_from_models_dir() for path conversion (#29674) This is part of the fs::path modernization series. That was also the opportunity to remove fs_list(). Signed-off-by: Adrien Gallouë

What's Changed Models run on MLX on Apple Silicon by default In this release, on Apple Silicon devices, model architectures supported by the MLX runtime will automatically run on MLX. ollama pull qwen

CUDA: Handle compute type for NVFP4 on cublass path (#29173) CUDA: Handle compute type for NVFP4 on cublass path Signed-off-by: ynankani ynankani@nvidia.com Use BF16 compute type for quantized models

model : re-enable -sm tensor for qwen4exp (#28569) #27941 disabled -sm tensor for qwen4exp because test-llama-archs asserted on the Meta device once the fixture carried a PLE layer: GGML_ASSERT(ggml_b

mtmd: cap max_image to ubatch for non_causal models (#29773) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/51957111 macOS/iOS: macOS Apple Silicon (arm64)

llama : add llama_prec_policy + model-driven W4A4 path (#24364) Rebase and update based on #26675 Signed-off-by: ynankani ynankani@nvidia.com CI failure fix(launh_bounds overload on HIP) and cleanup S

model : support Gemma4 DSpark draft backbone (#29226) dspark: add Gemma 4 draft support Add GGUF conversion and runtime support for full-attention and SWA Gemma 4 DSpark drafts, including tied output

chat : add dedicated Ling 3.0 (Bailing V3) parser (#28682) chat: add dedicated Ling 3.0 (Bailing V3) parser Ling 3.0 Flash templates pre-open the think block in the generation prompt, so the model nev

vocab : add ufakzeka pre-tokenizer (#29033) vocab : add ufakzeka pre-tokenizer vocab : move ufakzeka to the models list and regenerate the hash mapping Website: https://llama.app Attestations: https:/

model : extend Nemotron MTP support (#29018) first fix removed unnecessary declarations Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/48188464 macOS/iOS:

model : skip gate_up_exps if TENSOR_SKIP is set (#29014) required for qwen35moe if MTP tensors are fused but not loaded Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/a

models : move build_arch_graph() after graph() template specialization (#28934) Move build_arch_graph()'s function definitions after the graph and graph template specializations have been explicitly d

models : guard the expert FFN size fallback in nemotron-h against a zero divisor (#28779) The NextN/MTP tail loop derives the expert FFN size as n_ff/n_expert_used when expert_feed_forward_length give

hexagon: support for multi-device model split (aka row-split) (#28589) hex-row-split: add support for multi-device row spliting Co-authored-by: Max Krasnyansky maxk@qti.qualcomm.com hex-mdev: add work

spec: fix failed to decode mtmd chunk with DFlash (#28587) speculative: fix failed to decode mtmd chunk with DFlash When using DFlash w/ vision models, the drafter memory fails to allocate new tokens

model: fix granite3 moe unknown parameter count (#28632) Signed-off-by: Aaron Teo aaron.teo1@ibm.com Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/4620980

model : support Kimi-K3 recurrent-state rollback (#28466) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/45866131 macOS/iOS: macOS Apple Silicon (arm64) ma

misc : prevent RAM peaking at model loading stage (#27483) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/44939788 macOS/iOS: macOS Apple Silicon (arm64) m

model : load hparams.n_layer_nextn before n_layer() calls (#28159) load hparams.n_layer_nextn before n_layer() calls remove duplicate loads Website: https://llama.app Attestations: https://github.com/

dflash: pass missing NVFP4 scales to attention operations (#28000) DFlash2 NVFP4 draft models produced almost no accepted speculative tokens because the Q, K, V, and output projection scales were not

model: qwen4exp: reduce number of graph splits (#27880) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/43734155 macOS/iOS: macOS Apple Silicon (arm64) macO

models : support nanbeige4.2-3B (#27730) Co-authored-by: admin lizongqiang@kanzhun.com Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/43308396 macOS/iOS: m

model : simplify MiniMax-01 graph (#27790) Co-authored-by: Stanisław Szymczyk sszymczy@gmail.com Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/43387702 ma

Deepseek 4: -sm tensor (#26490) DSV4: sm tensor set coarser granularity for head splits fix dspark add model saving for dsv4 + allow dflash to return on specific device add comment about dsv4 seq_rm s

model : support MTP in GLM-4.5-Air (#26534) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/42447665 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Sili

model: add dots3-note (#27060) text: conversion init impl address review comments fix rope move to a new llama_kv_cache_dsa_iswa Website: https://llama.app Attestations: https://github.com/ggml-org/ll

model : support DSpark for bailingmoe3 (#27508) Website: https://llama.app Attestations: https://github.com/ggml-org/llama.cpp/attestations/42309942 macOS/iOS: macOS Apple Silicon (arm64) macOS Apple

server: add dedup-cache-models preset option (#27346) Website: https://llama.app macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabled) DISABLED macOS Intel (x64) iOS XCF

model : remove some ggml_concat (#27176) Co-authored-by: Xuan Son Nguyen son@huggingface.co Website: https://llama.app macOS/iOS: macOS Apple Silicon (arm64) macOS Apple Silicon (arm64, KleidiAI enabl

server: re-design yield_to_queue thread model (#27133) run common_speculative_process in worker swap worker <--> main thread design Website: https://llama.app macOS/iOS: macOS Apple Silicon (arm64) ma