AI News — Monday, July 13, 2026

10
OpenAI's GPT 5.6 is the 'preferred model' for Microsoft Copilot 365 amid breakup chatter

OpenAI confirms GPT 5.6 as the favored model for Microsoft Copilot 365, amidst ongoing speculation about potential tensions or a 'breakup' between the two AI giants.

TechCrunchindustry
9
OpenCoF: Learning to Reason Through Video Generation

Researchers introduce OpenCoF, a new framework that enables AI models to learn reasoning capabilities by generating videos, marking a significant step towards more intelligent video understanding.

Hugging Faceresearch
9
An AI agent startup just let its agent run its $100M fundraise

A new AI agent startup successfully closed a $100 million funding round, notably allowing its own AI agent to autonomously manage the entire fundraising process.

TechCrunchindustry
8
OpenAI is shutting down Atlas, but its AI browser ambitions are still growing

OpenAI is discontinuing its Atlas project, yet simultaneously signals continued and expanding ambitions in the development of AI-powered browser technologies.

TechCrunchindustry
8
Nous Research's NousCoder-14B is an open-source coding model landing right in the Claude Code moment

Nous Research has released NousCoder-14B, a new open-source coding model that positions itself as a strong competitor in the rapidly evolving landscape of AI-powered code generation, particularly against models like Claude Code.

VentureBeatopen-source
8
CineMobile: On-Device Image-to-Video Diffusion for Cinematic Camera Motion Generation

CineMobile introduces an on-device image-to-video diffusion model capable of generating cinematic camera motions, bringing advanced video creation to mobile platforms.

Hugging Faceresearch
7
Simple Benchmark Review: Ollama on Jetson Nano

A practical benchmark review evaluates the performance of Ollama, a tool for running large language models locally, on the resource-constrained Jetson Nano device.

Dev.toproduct
7
Linear Attention Architectures: Mechanisms, Trade-offs, and Cross-Layer Routing

This paper delves into the intricacies of linear attention architectures, analyzing their mechanisms, performance trade-offs, and novel cross-layer routing strategies for improved efficiency.

Hugging Faceresearch
7
RoboDojo: A Unified Sim-and-Real Benchmark for Comprehensive Evaluation of Generalist Robot Manipulation Policies

RoboDojo presents a new unified benchmark that bridges simulation and real-world environments to thoroughly evaluate generalist robot manipulation policies, advancing the field of robotics.

Hugging Faceresearch
6
Let an AI clear out your false positives without letting it hide a real bug

This article explores strategies for leveraging AI to filter out false positives in bug detection while ensuring that genuine software defects are not inadvertently concealed.

Dev.toproduct
6
Remember When It Matters: Proactive Memory Agent for Long-Horizon Agents

Researchers introduce a proactive memory agent designed to enhance the performance of long-horizon AI agents by selectively recalling relevant information when needed.

Hugging Faceresearch
6
Sparse Delta Memory: Scaling the State of Linear RNNs through Sparsity

This paper proposes Sparse Delta Memory, a method that leverages sparsity to efficiently scale the state of linear Recurrent Neural Networks, improving their capacity and performance.

Hugging Faceresearch
6
Automating the Design of Embodied Agent Architectures

This research focuses on automating the design process for embodied agent architectures, aiming to streamline the development of intelligent systems that interact with physical environments.

Hugging Faceresearch
6
InsightsTrack + Pulse: I taught Claude Desktop to read my web analytics (via MCP)

A developer successfully integrated Claude Desktop with web analytics tools (InsightsTrack + Pulse) using an MCP (Multi-modal Control Protocol) to enable the AI to interpret and act on data.

Dev.toproduct
5
Scalable Visual Pretraining for Language Intelligence

This paper explores scalable visual pretraining techniques designed to significantly enhance the language intelligence capabilities of AI models.

Hugging Faceresearch