MoonshotAI
Featured
Kimi-K3
Kimi K3 is an open-weight, 2.8T-parameter native multimodal agentic model with a 1M-token context window, designed for frontier coding, knowledge work, and reasoning tasks.
OpenCoF
OpenCoF is a framework that improves video generation models for reasoning by introducing a Chain-of-Frame (CoF) approach, a curated 17K video dataset, and reasoning token designs, achieving gains on multiple benchmarks.
Stars
73
7-day growth
No data
Forks
0
Open issues
2
License
No data
Last updated
2026-07-10
AI repository intelligence
FR-AI / ANALYSISWhy it is worth attention
It pioneers a new reasoning paradigm (Chain-of-Frame) distinct from Chain-of-Thought, provides a diverse reasoning video dataset with four curation pipelines, and demonstrates significant performance improvements on external video-reasoning benchmarks through data supervision and reasoning token designs.
Who it is for
- Researchers in video generation and multimodal reasoning
- AI practitioners working on reasoning in large models
- Developers of synthetic dataset generation pipelines
- Academics studying Chain-of-Thought alternatives
Use cases
- Improving video generation models for logical reasoning tasks
- Benchmarking video reasoning capabilities across multiple tasks
- Generating synthetic reasoning videos via instance-based, expert-guided, procedural, and repurposing pipelines
- Exploring token-level reasoning mechanisms for video generation
Strengths
- Introduces a novel Chain-of-Frame reasoning paradigm that leverages temporal frame connections
- Curates a large-scale reasoning video dataset (OpenCoF-17K) spanning 11 task families with four complementary pipelines
- Wan-CoF model outperforms baseline Wan2.2 on four external video-reasoning benchmarks
- Reasoning token designs (visual and textual) provide further gains by capturing low-level and high-level cues
Considerations
- Code, dataset, and model checkpoints are not yet publicly available (pending internal review)
- The framework is built on a single base model (Wan2.2), so generality to other architectures is unverified
- Evaluation is limited to four specific benchmarks; real-world applicability may require further testing
README quick start
OpenCoF: Learning to Reason Through Video Generation
Official repository for the paper "OpenCoF: Learning to Reason Through Video Generation".
💥 News
- [2026.07] Code, dataset, and model release is pending internal company review.
👀 About OpenCoF
Reasoning has become a core capability for large models, especially when reliable decisions require understanding logical consequences. Recent video generation models offer a reasoning path distinct from previous Chain-of-Thought (CoT): reasoning can unfold through temporally connected frames, known as Chain-of-Frame (CoF) reasoning. However, existing video generators are primarily trained on general video corpora, still lacking diverse supervision and dedicated designs for CoF reasoning.
To address this gap, we introduce OpenCoF, a framework built around OpenCoF-17K, a reasoning video dataset of 17,312 videos spanning 11 task families, curated through four complementary pipelines: instance-based rendering, expert-guided rendering, procedural scene synthesis, and external video repurposing.
We use OpenCoF-17K to fine-tune Wan2.2-I2V-A14B into Wan-CoF, which achieves considerable gains over the baseline across four external video-reasoning benchmarks (MME-CoF, Gen-ViRe, VIPER, RULER-Bench) purely from data supervision. Building on this, we further explore two complementary reasoning-token designs, Visual Reasoning Tokens (vt) and Textual Reasoning Tokens (tt), which respectively capture low-level visual cues and high-level semantic priors, yielding Wan-CoFvt and Wan-CoFtt with further gains.
Our results suggest that stronger video reasoning requires both broad temporal supervision and explicit mechanisms for organizing intermediate reasoning state.
🚧 Code, Model & Dataset
The code, model checkpoints, and dataset are currently going through internal company review before public release.
- Code — Coming soon
- OpenCoF-17K dataset — Coming soon
- Wan-CoF model checkpoints — Coming soon
📖 Citation
If you find OpenCoF useful for your research, please consider citing our paper:
@article{chen2026opencof,
title = {OpenCoF: Learning to Reason Through Video Generation},
autho
Description
OpenCoF: Learning to Reason Through Video Generation
Related repositories
Similar projects matched by category, topics, and programming language.
xuchonglang
Featured
investing-for-beginners
A structured investing guide for Chinese beginners covering US stocks, options, and cryptocurrency, with focus on foundational concepts and risk awareness.
Krishnagangwal
Featured
CS-Fundamentals
A curated collection of Computer Science fundamentals (PDFs, notes, cheatsheets, interview question banks) for placement preparation, covering seven core subjects plus general resources.