github llama.cpp · 1h ago · inference

llama.cpp b10254 adds DeepSeek V4 Flash 0731 template and updates V4 templates

from English

The llama.cpp project released version b10254, introducing a new chat template for the DeepSeek V4 Flash 0731 model alongside updates to existing DeepSeek V4 templates.

Aligns DeepSeek V4 templates with official encoders while excluding parser behavior changes.
Defaults to dropping thinking history for DeepSeek V4 unless preserve_reasoning is requested or tools are present.
Adds structured output response-format instructions and passes schemas into template rendering.
Introduces a separate Flash 0731 template to handle updated high and max reasoning effort mappings.
Includes fixes for DSML parser tool call separator consumption and deepseek v4 0731 template selection.

This update ensures compatibility with the latest DeepSeek model specifications and improves structured output handling.

Importance 1/3 Trust 1/3 llama.cpp Inference efficiency Local inference

Read original

github llama.cpp · 4h ago

llama.cpp b10251 adds MTP support for GLM-4.7-Flash

The llama.cpp project has released build b10251, which introduces Multi-Token Prediction (MTP) support for the GLM-4.7-Flash model.

github llama.cpp · 22h ago

llama.cpp b10238 adds MTP support for Qwen3-Next

The llama.cpp project released build b10238, which introduces Multi-Token Prediction (MTP) support for the Qwen3-Next model family. This update includes specific implementation changes to handle MTP layers and fixes related to Python type checking within the codebase.

github llama.cpp · 23h ago · 5 views

llama.cpp b10237 adds MTP support for DeepSeek V3.2

The llama.cpp project released build b10237, which introduces Multi-Token Prediction (MTP) support for the DeepSeek V3.2 model.

github llama.cpp · 2d ago · 7 views

llama.cpp b10232 implements DeepSeek V4 hyper-connections in Metal

The llama.cpp project released build b10232, which introduces support for DeepSeek V4 hyper-connections via the Metal backend. This update adds implementation of GGML_OP_DSV4_HC_COMB, GGML_OP_DSV4_HC_PRE, and GGML_OP_DSV4_HC_POST operations.

github llama.cpp · 3d ago · 9 views

llama.cpp b10219 persists reasoning_content in chat history

The llama.cpp project released version b10219, which includes a fix for the CLI to persist reasoning content within the chat history. Previously, while `llama-cli` collected reasoning from the stream for display, it only stored assistant content in messages, preventing the `--reasoning-preserve` flag from re-injecting prior thoughts on later turns.