ggml-org/llama.cpp

Feature Request: support the multi-modal Bagel model

Open

#17861 opened on Dec 8, 2025

View on GitHub
 (1 comment) (1 reaction) (0 assignees)C++ (110,169 stars) (18,202 forks)batch import
Review Complexity : Highenhancementhelp wantedmodelmtmd

Description

Prerequisites

  • I am running the latest code. Mention the version if possible as well.
  • I carefully followed the README.md.
  • I searched using keywords relevant to my issue to make sure that I am creating a new issue that is not already open (or closed).
  • I reviewed the Discussions, and have a new and useful enhancement to share.

Feature Description

support multi-modal Bagel

Motivation

https://github.com/bytedance-seed/BAGEL https://github.com/stepfun-ai/Step1X-Edit

Possible Implementation

No response

Contributor guide

Tech stack
cpp
Domain
aibackend
Issue type
feature
DifficultyEstimated implementation difficulty for a new contributor, from 1 for very small changes to 5 for expert-level work.
5
Estimated timeA rough time range for an experienced contributor to investigate, implement, test, and prepare a pull request.
over 1 week
Activity statusHow available the issue appears right now: fresh, active, stale, blocked, or waiting on maintainer input.
fresh
ClarityHow clearly the issue explains the expected change, acceptance criteria, and next step.
unclear
Prerequisites
C++llama.cpp architecturemulti modal model conceptsimage processing basics
Newbie friendlinessA 1-100 score estimating how approachable this issue is for first-time contributors.
10
Research direction
Investigate existing multi modal support in llama.cpp (e.g., LLaVA or similar models) to understand the codebase structure. Review the repositories linked in the issue (BAGEL and Step1X Edit) to understand the model architecture and required changes. Look at files like `llama.cpp` and `llama.h` for inference code, and examine any existing multi modal implementation patterns. Engage with maintainers in the issue comments to clarify the expected integration approach.