OpenBMB/MiniCPM-V
27.4
Weak · 16 November 2026
26k
lines of production code
Python
with JavaScript
5
measurements over time
What this system is
Features
Add MiniCPM-V 2.6 and MiniCPM-o 2.6 evaluation kit
The \eval\_mm\ directory now includes a complete evaluation environment for MiniCPM-V 2.6 and MiniCPM-o 2.6, featuring the \vlmevalkit\ package (version 0.2rc1) for benchmarking on OpenCompass datasets (e.g., MMMU, MathVista, MME) and the \vqaeval\ suite for VQA tasks (TextVQA, DocVQA). This update provides the necessary inference scripts, configuration files, and API wrappers to run these specific model versions against standard multimodal benchmarks.
_eval\mm · high confidence
Add MiniCPM-o 2.6 web demos and model server
New demo files have been added for the MiniCPM-o 2.6 model, including a Gradio-based chatbot (\chatbot\_web\_demo\_o2.6.py\), a FastAPI model server (\model\_server.py\), and a Vue.js web client (\web\_server/\). These components enable users to interact with the MiniCPM-o 2.6 model via a web interface, supporting multimodal inputs and real-time streaming responses.
_web\demos · high confidence
Add evaluation support for MiniCPM-V and MiniCPM-Llama3-V models
The evaluation framework now includes specific implementations for the MiniCPM-V and MiniCPM-Llama3-V-2\_5 vision-language models. This change adds the \minicpm\_v.py\ module, which defines \MiniCPM\_V\ and \MiniCPM\_Llama3\_V\ classes to handle model loading, prompt construction, and inference. The implementation supports various dataset types (MCQ, VQA, Y/N) and specific benchmarks (such as MME, HallusionBench, and MMMU) with tailored system prompts and generation parameters, enabling users to evaluate these specific model architectures within the vlmeval suite.
vlmeval · high confidence
New assets added for MiniCPM-O branding and language support
The repository now includes a new HTML logo asset for MiniCPM-O and a comprehensive Markdown file listing the 40+ languages supported by MiniCPM-Llama-V 2.5. Additionally, a star history SVG chart has been added to visualize repository growth, and a temporary placeholder file was created within the MiniCPM-V 2.5 assets directory.
assets · high confidence
New finetuning script and VAD utility for MiniCPM-O 2.6
Users can now fine-tune MiniCPM models using the new \finetune/finetune.py\ script, which supports supervised fine-tuning with LoRA, DeepSpeed ZeRO-3, and configurable vision/LLM tuning. Additionally, the \web\_demos/minicpm-o\_2.6/vad\_utils.py\ module provides voice activity detection utilities for the MiniCPM-O 2.6 web demo, including speech chunking and silence detection.
python · high confidence
New local chatbot demo for MiniCPM-o 2.6
The web server for the local chatbot demo has been updated to support MiniCPM-o 2.6. This introduces a new Vue-based frontend interface featuring a call header, a countdown timer for call duration, and a model output area that displays text and audio responses. The demo includes a model configuration panel allowing users to adjust settings such as audio interruption, video quality, VAD threshold, and voice prompts. It also supports voice interaction features like voice cloning, timbre selection, and user feedback (like/dislike) with comments. The backend API endpoints have been updated to handle streaming messages, stopping messages, uploading audio files, and submitting feedback.
_web\_demos/minicpm-o\_2.6/web\server · high confidence
New official finetuning scripts and infrastructure for MiniCPM-V and MiniCPM-o models
This change introduces a new \finetune\ module providing official scripts and configuration for fine-tuning MiniCPM-V 4.0, MiniCPM-o 2.6, MiniCPM-V 2.6, MiniCPM-Llama3-V 2.5, and MiniCPM-V 2.0. It includes \finetune\_ds.sh\ for full-parameter fine-tuning and \finetune\_lora.sh\ for LoRA-based tuning, both supporting multi-image inputs and DeepSpeed Zero-2/Zero-3 optimization. The module adds a custom \CPMTrainer\ to handle model-specific loss computation and a \SupervisedDataset\ class that manages tokenization, image preprocessing, and conversation formatting for supervised fine-tuning.
finetune · high confidence
Behavioural changes
Disabled verbose logging for vision tower weight loading
The OmniLMM model no longer prints a log message indicating that vision tower weights are being skipped during initialization. This change suppresses console output, resulting in a cleaner execution log for users running the model.
omnilmm · high confidence
Repository rebrands from OmniLMM to MiniCPM-V and MiniCPM-o with updated model support
The project has transitioned its primary focus from the OmniLMM series to the MiniCPM-V and MiniCPM-o model families. The documentation (README.md and README\_zh.md) has been completely rewritten to highlight MiniCPM-V 4.6 and MiniCPM-o 4.5, replacing previous references to OmniLMM-3B and OmniLMM-12B. Correspondingly, the Python inference script (chat.py) has been refactored: the main controller class is renamed from OmniLMMChat to MiniCPMVChat, and the underlying model classes have been updated to support MiniCPM-V, MiniCPM-Llama3-V 2.5, and MiniCPM-V 2.6, including new logic for multi-GPU inference in version 2.6.
(repo-wide) · high confidence
Dependencies
Updated Python dependencies and added new evaluation requirements files
This change updates core Python dependencies for the main project, upgrading PyTorch to 2.1.2, Transformers to 4.40.0, and Gradio to 4.41.0, while removing the base\_utils dependency. It also introduces three new requirements files: eval\_mm/vlmevalkit/requirements.txt and eval\_mm/vqaeval/requirements.txt for evaluation toolkits, and requirements\_o2.6.txt for the MiniCPM-o 2.6 demo, which pins specific versions for libraries like accelerate, sentencepiece, and gradio to ensure compatibility with the new model version.
(dependencies) · high confidence
Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.
How this codebase got here
Score
- CAI 30 → 27 (-2.6)
- Rubric changed (rubric-2026.11.2 → rubric-2026.11.4) — scores are not directly comparable.
Lenses
- Code Health 29 → 18 (-11.1)
- Architecture 94 → 95 (+0.6)
- Maturity 63 → 63 (+0.1)
- Readiness 16 → 16 (+0.0)
- Security 76 → 76 (+0.0)
- Accessibility 30 → 31 (+1.0)
- Performance 100 → 100 (+0.0)
Resolved (81)
- Critical CVE: [GHSA redacted] (finetune/requirements.txt)
- Critical CVE: [GHSA redacted] (finetune/requirements.txt)
- Critical CVE: [GHSA redacted] (requirements.txt)
- Critical CVE: [GHSA redacted] (requirements_o2.6.txt)
- Critical CVE: [GHSA redacted] (finetune/requirements.txt)
- Critical CVE: [GHSA redacted] (web_demos/minicpm-o_2.6/web_server/pnpm-lock.yaml)
- Critical CVE: [GHSA redacted] (eval_mm/vqaeval/requirements.txt)
- Critical CVE: [GHSA redacted] (requirements.txt)
- Critical CVE: [GHSA redacted] (requirements_o2.6.txt)
- Critical CVE: [GHSA redacted] (web_demos/minicpm-o_2.6/web_server/pnpm-lock.yaml)
- Duplicated block (10 lines × 2) (eval_mm/vlmevalkit/vlmeval/dataset/cmmmu.py)
- Duplicated block (12–13 lines × 2) (eval_mm/vlmevalkit/vlmeval/dataset/utils/longvideobench.py)
- Duplicated block (13–19 lines × 2) (eval_mm/vlmevalkit/vlmeval/dataset/image_vqa.py)
- Duplicated block (16 lines × 2) (eval_mm/vlmevalkit/vlmeval/dataset/cgbench.py)
- Duplicated block (17 lines × 2) (eval_mm/vlmevalkit/vlmeval/vlm/minicpm_v.py)
- Duplicated block (18 lines × 3) (eval_mm/vlmevalkit/vlmeval/dataset/dude.py)
- Duplicated block (21 lines × 2) (eval_mm/vlmevalkit/vlmeval/dataset/cgbench.py)
- Duplicated block (26–32 lines × 2) (eval_mm/vlmevalkit/vlmeval/dataset/image_vqa.py)
- Duplicated block (32 lines × 2) (eval_mm/vlmevalkit/vlmeval/dataset/cgbench.py)
- Duplicated block (35 lines × 2) (eval_mm/vlmevalkit/vlmeval/vlm/minicpm_v.py)
- …and 61 more
New (69)
- Critical CVE: [CVE redacted] (web_demos/minicpm-o_2.6/web_server/pnpm-lock.yaml)
- Critical CVE: [CVE redacted] (web_demos/minicpm-o_2.6/web_server/pnpm-lock.yaml)
- Duplicated block (106 lines × 2) (eval_mm/vlmevalkit/vlmeval/dataset/cgbench.py)
- Duplicated block (44 lines × 2) (eval_mm/vlmevalkit/vlmeval/dataset/cgbench.py)
- Duplicated block (6–7 lines × 2) (eval_mm/vlmevalkit/vlmeval/dataset/image_mcq.py)
- Edited copies of one member (3 members, 11+ corresponding lines) (eval_mm/vlmevalkit/vlmeval/dataset/dude.py)
- Edited copies of one member (3 members, 15+ corresponding lines) (eval_mm/vlmevalkit/vlmeval/dataset/mlvu.py)
- Edited copies of one member (3 members, 15+ corresponding lines) (eval_mm/vlmevalkit/vlmeval/dataset/tempcompass.py)
- Edited copies of one member (3 members, 20+ corresponding lines) (eval_mm/vlmevalkit/vlmeval/dataset/dynamath.py)
- Edited copies of one member (3 members, 47+ corresponding lines) (eval_mm/vlmevalkit/vlmeval/dataset/dude.py)
- Edited copies of one member (3 members, 48+ corresponding lines) (eval_mm/vlmevalkit/vlmeval/dataset/tempcompass.py)
- Edited copies of one member (3 members, 48+ corresponding lines) (eval_mm/vlmevalkit/vlmeval/vlm/minicpm_v.py)
- Edited copies of one member (3 members, 58+ corresponding lines) (eval_mm/vlmevalkit/vlmeval/dataset/mlvu.py)
- Edited copies of one member (3 members, 67+ corresponding lines) (eval_mm/vlmevalkit/vlmeval/dataset/tempcompass.py)
- Edited copies of one member (4 members, 5+ corresponding lines) (omnilmm/model/utils.py)
- Edited copy of a member (11 corresponding lines) (eval_mm/vlmevalkit/vlmeval/dataset/mvbench.py)
- Edited copy of a member (12 corresponding lines) (eval_mm/vlmevalkit/vlmeval/dataset/mmmath.py)
- Edited copy of a member (12 corresponding lines) (eval_mm/vlmevalkit/vlmeval/dataset/mmmath.py)
- Edited copy of a member (12 corresponding lines) (eval_mm/vlmevalkit/vlmeval/vlm/base.py)
- Edited copy of a member (13 corresponding lines) (eval_mm/vlmevalkit/vlmeval/dataset/mvbench.py)
- …and 49 more
Changes since last survey
- 1 commits — 1 feature/other, 0 fixes
By area
- (root) — 1 commit
Notable commits
- change: docs: update docs link to openbmb.cn/docs
Architecture
- Unchanged — 0 containers · 2 contexts · 0 edges
Written by cca.canine.dev from the codebase's own history, inside the signed delivery this page is composed from.
Survey your own repository
OpenBMB/MiniCPM-V was measured the same way every project in this corpus was: the same rubric, at a pinned commit, with the result published in full. Point a surveyor at a repository you know and see whether you agree with it.
About this page
- The score is its most recent published measurement, taken on 16 November 2026 at a pinned commit. It is not a live figure and does not change until the project is measured again.
- Measured at commit 4dbbb6c58b6f8a004d6689028417296719f0831a — the exact code this score is about.
- Scored under rubric-2026.11.4 — the same rubric and the same method as every other entry in this index.
- Measured by Canine Code Analyzer (cca.canine.dev) using cca measure-2026.11.73+preprod-52de90b74216.