MCP App StoreLe meilleur de l'écosystème IA et MCP, sélectionné chaque jour.
Sources
OpenAI's year-end retrospective for developers covering the biggest model, API, and platform milestones of 2025 — from the Responses API launch to Codex, GPT-5.4, Realtime API, and the shift toward production-grade agentic systems. A useful reference for understanding where the platform stands heading into 2026.
ServiceNow AI presents AprielGuard, a new safety guardrail designed to improve adversarial robustness in LLM systems. The framework focuses on preventing malicious prompts and ensuring safe model outputs in production environments.
OpenAI released new audio model snapshots and expanded Custom Voices access to production voice apps. The updates improve speech quality and latency in the Realtime API and make it easier to deploy branded voices in customer-facing products. Key update for teams building voice agents or voice-first interfaces.
OpenAI is utilizing automated red teaming and reinforcement learning to proactively identify and patch prompt injection vulnerabilities in ChatGPT Atlas. This is critical for the stability and security of AI agents interacting with browser environments.
The MCP Transport Working Group outlines plans to evolve beyond Streamable HTTP for enterprise-scale remote deployments — addressing the stateful connection bottleneck that makes load balancing and managed services difficult. Proposes a direction that preserves backward compatibility while enabling stateless, horizontally-scalable MCP server deployments. Critical reading for teams deploying MCP at scale.
OpenAI presents a framework for monitoring internal chain-of-thought reasoning in AI models. The research indicates that observing internal reasoning is more effective for scalable control than monitoring final outputs alone.
OpenAI released a system card addendum for GPT-5.2-Codex, detailing the technical specifications and safety evaluations for the coding-specialized model variant.
OpenAI released GPT-5.2-Codex, featuring advanced long-horizon reasoning and large-scale code transformation capabilities. It specifically targets complex software engineering tasks and cybersecurity enhancements.
OpenAI has launched a directory for ChatGPT apps, allowing developers to submit their tools for review and publication. An updated Apps SDK and new guidelines are provided to help build chat-native experiences with real-world actions.
OpenAI introduces a new evaluation framework to measure AI's impact on biological wet-lab research. The study utilizes GPT-5 to optimize molecular cloning protocols, highlighting both potential efficiency gains and safety risks.
OpenAI has launched an upgraded image generation model for ChatGPT and the API as GPT-Image-1.5. The new model offers faster generation speeds (up to 4x) and improved precision in edits and consistency.
IBM Research introduces CUGA on Hugging Face, a framework for creating and sharing configurable AI agents. This initiative aims to democratize agent development by providing a standardized way to define and deploy agentic behaviors.
OpenAI detailing the rapid development of Sora for Android using Codex. Highlights how AI-assisted planning and parallel coding accelerated the shipping timeline to just 28 days.
llama.cpp introduces new model management capabilities, improving how local LLMs are handled. This is a key update for developers building edge-AI applications and local inference pipelines.
Introduction of GPT-5.2, OpenAI's most capable model for mathematics and science. The model achieves new state-of-the-art results on benchmarks like GPQA Diamond and FrontierMath.
Codex is open-sourcing its AI models to promote transparency and democratize access to high-quality code generation tools. This move supports the open-science movement in AI development.
OpenAI has released GPT-5.2, a frontier model optimized for professional work. It features significant improvements in reasoning, coding, and vision to support more reliable agentic workflows via API and ChatGPT.
OpenAI and Disney have partnered to bring over 200 iconic characters to Sora for fan-inspired short videos. The deal also includes the adoption of ChatGPT Enterprise and the OpenAI API across Disney's company.
Anthropic donated MCP to the newly formed Agentic AI Foundation — a directed fund under the Linux Foundation — ensuring vendor-neutral governance for the protocol's future. MCP becomes a founding project of the foundation alongside other agentic AI standards. A major governance milestone that signals MCP's transition from an Anthropic project to a true open standard.
OpenAI has co-founded the Agentic AI Foundation under the Linux Foundation, donating the AGENTS.md standard. This initiative aims to establish open, interoperable standards for the safe development and deployment of agentic AI systems.