Latest stories
VideoGen-Agent Trains AI to Orchestrate Video Tools, Boosting Generation Scores by 19 Points
Hugging Face researchers introduce VideoGen-Agent, a multimodal agent trained with multitask agentic reinforcement learning to coo...
Tencent's GameHorizon Suite Offers a New Yardstick for AI Gameplay Skills
A new research suite from Tencent ARC introduces 5,000 hours of annotated AAA gameplay data and a benchmark that evaluates 47 mode...
RoboDawn Lets Vision-Language Models Control Robots Without Task-Specific Training
A new Hugging Face paper introduces RoboDawn, an interface that lets an agentic VLM control a robot through discrete translation,...
WorldCrafter Builds 3D-Aware Memory for Consistent Video World Models
A new video world model, WorldCrafter, learns a camera-queryable implicit 3D-aware memory that keeps long-horizon scene exploratio...
Google Research's RRSI Cuts Agent Harness Overfitting in Recursive Self-Improvement
A new paper introduces RRSI, a regularized approach to recursively self-improving LLM agent harnesses that favors reusable mechani...
Higgsfield AI Ships New Video Features in a Day Using GPT-6 Astra
Higgsfield AI used OpenAI's GPT-6 Astra to launch new video features in a single day, making video ad creation easier for small bu...
OpenAI Forms Advisory Group on Mathematics and AI
OpenAI is collaborating with an independent advisory group to guide the review and communication of emerging AI results in mathema...
OpenAI Calls for Coordinated Global AI Standards
OpenAI has outlined a path toward shared global AI standards, urging coordinated evaluation, reporting, and governance to strength...
OpenAI Academy Adds New Learning Paths for AI Skills
OpenAI has expanded its Academy with new learning paths aimed at employees, developers, leaders, educators, and students who want...
CADWorld Benchmark Tests AI Agents on Long-Horizon FreeCAD Workflows
Hugging Face paper introduces CADWorld, a 200-task benchmark for computer-use agents in mechanical CAD, where the best of seven ag...
AI Reviews Train AI Reviewers: Study Warns of 'Scientific-Judgment Collapse'
A new Hugging Face paper shows that training reviewer models on AI-generated reviews compresses rating distributions and reduces s...
New HEAL Method Targets Hallucinations in Multimodal LLMs via 'Synergy Heads'
A new paper introduces HEAL, a head-level information disentanglement and calibration method that identifies and reduces hallucina...