PerceptionBench: Evaluating Atomic Visual Perception in Multimodal Large Language Models

Open Frontier Intelligence

Open source Ghostty-based macOS terminal with vertical tabs and notifications for AI coding agents. Built for multitasking, organization, and programmability.

WorldReasonBench: Human-Aligned Stress Testing of Video Generators as Future World-State Predictors

Beyond SFT-to-RL: Pre-alignment via Black-BoxOn-Policy Distillation for Multimodal RL

🔥[VLDB’26] Synthesizing Data Agent Trajectories via Execution-Grounded Tree Search

ParaVT: Taming the Tool Prior Paradox for Parallel Tool Use in Agentic Video Reinforcement Learning

[Roadmap] Visual Generation in the New Era: An Evolution from Atomic Mapping to Agentic World Modeling

V-CAST: Video Curvature-Aware Spatio-Temporal Pruning for Efficient Video Large Language Models

mini-claude-code: A minimal, readable Python re-implementation of Claude Code

VTC-R1: Vision-Text Compression for Efficient Long-Context Reasoning.

A fork to add multimodal model training to open-r1

🦦 Otter, a multi-modal model based on OpenFlamingo (open-sourced version of DeepMind's Flamingo), trained on MIMIC-IT and showcasing improved instruction-following and in-context learning ability.

One-for-All Multimodal Evaluation Toolkit Across Text, Image, Video, and Audio Tasks

本人的科研经验

The RL Bridge for LLM-based Agent Applications. Made Simple & Flexible.

The open source coding agent.

Transforms complex documents like PDFs and Office docs into LLM-ready markdown/JSON for your Agentic workflows.

DeepResearchEval: An Automated Framework for Deep Research Task Construction and Agentic Evaluation.

EditReward: A Human-Aligned Reward Model for Instruction-Guided Image Editing [ICLR 2026]

Incentivizing "Thinking with Long Videos" via Native Tool Calling

Multi-modal Critical Thinking Agent Framework for Complex Visual Reasoning

A minimal, educational HEVC (H.265) encoder written in Python.

🔥[ICDE'26] Official repository for the paper "CARROT: A Learned Cost-Constrained Retrieval Optimization System for RAG".

Official implementation of MATPO: Multi-Agent Tool-Integrated Policy Optimization.