2/6/2026
From Features to Actions: Explainability in Traditional and Agentic AI Systems
Research on explainability methods for agentic AI systems that operate over multi-step trajectories, extending beyond single-prediction interpretability.
2/5/2026
Sparse Video Generation Propels Real-World Beyond-the-View Vision-Language Navigation
Sparse video generation framework enables vision-language navigation agents to navigate unknown environments with minimal high-level instructions via beyond-the-view reasoning.
2/5/2026
Unveiling Implicit Advantage Symmetry: Why GRPO Struggles with Exploration and Difficulty Adaptation
Analysis of GRPO reinforcement learning limitations in LLM reasoning due to implicit advantage symmetry; proposes improvements for exploration and difficulty adaptation.
2/5/2026
GPT-5 lowers the cost of cell-free protein synthesis
Autonomous lab combining GPT-5 with Ginkgo Bioworks' automation reduced cell-free protein synthesis costs 40% via closed-loop experimentation.
2/5/2026
Introducing GPT-5.3-Codex
GPT-5.3-Codex is a Codex-native agent for long-horizon technical work pairing coding performance with general reasoning.
2/5/2026
GPT-5.3-Codex System Card
GPT-5.3-Codex system card detailing capabilities of an agentic coding model combining frontier coding performance with reasoning abilities.
2/4/2026
Unlocking the Codex harness: how we built the App Server
Technical guide to embedding Codex agent via App Server using bidirectional JSON-RPC API with streaming, tool use, and diffs.
2/4/2026
GeneralVLA: Generalizable Vision-Language-Action Models with Knowledge-Guided Trajectory Planning
GeneralVLA: vision-language-action model with knowledge-guided trajectory planning to improve zero-shot generalization in robotic control.
2/4/2026
BPDQ: Bit-Plane Decomposition Quantization on a Variable Grid for Large Language Models
BPDQ: bit-plane decomposition quantization with variable grid for efficient 2-3 bit LLM inference under memory constraints.
2/3/2026
Reasoning Cache: Continual Improvement Over Long Horizons via Short-Horizon RL
Reasoning Cache (RC) algorithm enables LLMs to improve over long horizons via test-time adaptation and RL, improving extrapolation beyond training distribution.
2/3/2026
An image is worth NxN words | Diffusion Transformers (ViT, DiT, MMDiT)
Investigates computational efficiency advantages of diffusion-based language models versus standard approaches.
2/3/2026
Quantized Evolution Strategies: High-precision Fine-tuning of Quantized LLMs at Low-precision Cost
Novel method for fine-tuning quantized LLMs using evolution strategies instead of backpropagation, enabling high-precision adaptation on discrete, non-differentiable parameter spaces.
2/2/2026
TIC-VLA: A Think-in-Control Vision-Language-Action Model for Robot Navigation in Dynamic Environments
TIC-VLA framework for robot navigation that models delayed semantic reasoning in vision-language-action models for real-time control in dynamic environments.
2/2/2026
ECHO-2: A Large-Scale Distributed Rollout Framework for Cost-Efficient Reinforcement Learning
ECHO-2 distributed RL framework for LLM post-training with remote inference workers, addressing cost efficiency and policy coordination challenges.
1/29/2026
Inside OpenAI’s in-house data agent
OpenAI's internal data agent architecture using GPT-5, Codex, and memory systems to reason over large datasets and deliver insights.
1/29/2026
Retiring GPT-4o, GPT-4.1, GPT-4.1 mini, and OpenAI o4-mini in ChatGPT
Announcement of model retirements: GPT-4o, GPT-4.1, GPT-4.1 mini, and o4-mini retiring from ChatGPT on February 13, 2026.
1/29/2026
Taisei Corporation shapes the next generation of talent with ChatGPT
Taisei Corporation scales ChatGPT Enterprise across HR and global construction operations for talent development.
1/28/2026
Keeping your data safe when an AI agent clicks a link
OpenAI details data protection mechanisms for AI agents, preventing URL-based exfiltration and prompt injection attacks.
1/27/2026
Powering tax donations with AI powered personalized recommendations
TRUSTBANK partners with Recursive to build Choice AI, a multi-agent system using OpenAI models for personalized donation recommendations.
1/26/2026
Unlock data from your files with Agentic Document Extraction
Opinion article debating whether prompt-driven development qualifies as legitimate programming practice.
1/23/2026
Document AI: How Agents add the brain to OCR's eyes
Technical guide on implementing agentic workflows for automated document processing and data extraction.
1/23/2026
Unrolling the Codex agent loop
Technical deep dive into Codex agent architecture, explaining orchestration of models, tools, prompts, and performance via Responses API.
1/22/2026
The Physics of Diffusion Models
Technical overview of diffusion transformers including vision transformers and diffusion transformer variants.
1/22/2026
Scaling PostgreSQL to power 800 million ChatGPT users
Technical case study on scaling PostgreSQL to handle millions of queries per second for ChatGPT using replicas, caching, and workload isolation.
1/22/2026
Inside Praktika's conversational approach to language learning
Praktika builds adaptive AI language tutors using GPT-4.1 and GPT-5.2 for personalized lessons and progress tracking.
1/22/2026
Inside GPT-5 for Work: How Businesses Use GPT-5
Data-driven report analyzing worker adoption patterns and use cases of ChatGPT across industries and departments.
1/21/2026
How Higgsfield turns simple ideas into cinematic social videos
Higgsfield uses GPT-4.1, GPT-5, and Sora 2 to generate cinematic social videos from simple user inputs.
1/20/2026
Cisco and OpenAI redefine enterprise engineering with AI agents
Cisco and OpenAI's Codex agent embeds AI into engineering workflows to automate builds, defect fixes, and enable AI-native development.
1/17/2026
Text Diffusion: A new LLM paradigm
Explores text diffusion models as emerging alternative paradigm for language model development.
1/13/2026
Veo 3.1 Ingredients to Video: More consistency, creativity and control
Veo 3.1 video generation model improves consistency and control with vertical video support
1/13/2026
Zenken boosts a lean sales team with ChatGPT Enterprise
Zenken deploys ChatGPT Enterprise company-wide, improving sales performance, preparation time, and proposal success rates.
1/8/2026
Netomi’s lessons for scaling agentic systems into the enterprise
Netomi scales enterprise AI agents using GPT-4.1 and GPT-5.2 with concurrency, governance, and multi-step reasoning for production workflows.
1/7/2026
How Tolan builds voice-first AI with GPT-5.1
Tolan builds voice-first AI companion with GPT-5.1 using low-latency responses, real-time context reconstruction, and memory-driven personalities.
12/22/2025
Continuously hardening ChatGPT Atlas against prompt injection
OpenAI hardens ChatGPT Atlas against prompt injection attacks using automated red teaming with reinforcement learning for proactive vulnerability discovery.
12/18/2025
Introducing GPT-5.2-Codex
OpenAI releases GPT-5.2-Codex, an advanced coding model with long-horizon reasoning, large-scale code transformations, and cybersecurity capabilities.
12/18/2025
Introducing GPT-5.2-Codex
OpenAI releases GPT-5.2-Codex, an advanced coding model with long-horizon reasoning, large-scale code transformations, and cybersecurity capabilities.
12/18/2025
Addendum to GPT-5.2 System Card: GPT-5.2-Codex
Safety documentation for GPT-5.2-Codex covering model-level and product-level mitigations for coding tasks and agent deployment.
12/17/2025
Developers can now submit apps to ChatGPT
ChatGPT app marketplace launched; developers can submit apps for review with updated SDK and guidelines for chat-native integrations.
12/17/2025
The state of enterprise AI
Overview of enterprise AI adoption, challenges, and potential for funding accessible AI. Discusses organizational AI implementation but lacks technical depth.
12/16/2025
Gemma Scope 2: helping the AI safety community deepen understanding of complex language model behavior
Gemma Scope 2 open-source interpretability tools release for Gemma 3 models to study language model behavior
12/11/2025
Advancing science and math with GPT-5.2
GPT-5.2 achieves state-of-the-art on math/science benchmarks (GPQA Diamond, FrontierMath) with applications to research and mathematical proofs.
12/11/2025
Introducing GPT-5.2
GPT-5.2 announced as advanced frontier model with reasoning, long-context, coding, and vision capabilities for agentic workflows.
12/11/2025
Update to GPT-5 System Card: GPT-5.2
Safety mitigation documentation for GPT-5.2 model family covering training data sources and safety approaches.
12/9/2025
FACTS Benchmark Suite: Systematically evaluating the factuality of large language models
FACTS Benchmark Suite provides systematic evaluation framework for measuring factuality in large language models
12/9/2025
OpenAI co-founds Agentic AI Foundation, donates AGENTS.md
OpenAI co-founds Agentic AI Foundation under Linux Foundation, donates AGENTS.md standard for safe, interoperable agentic AI systems.
12/9/2025
Launching our first OpenAI Certifications courses
OpenAI launches certification courses and AI Foundations program to build real-world AI skills and career preparation.
12/9/2025
Building AI fluency at scale with ChatGPT Enterprise
Commonwealth Bank deploys ChatGPT Enterprise to 50,000 employees to build AI fluency and improve customer service and fraud detection.
12/3/2025
How confessions can keep language models honest
OpenAI researchers develop 'confessions' training method that teaches models to acknowledge mistakes, improving transparency and trustworthiness.
12/1/2025
Inside Mirakl's agentic commerce vision
Mirakl uses AI agents and ChatGPT Enterprise for commerce optimization, improving documentation and customer support with agent-native architecture.
12/1/2025
Accenture and OpenAI accelerate enterprise AI success
Accenture and OpenAI partnership to deploy agentic AI in enterprises, with ChatGPT Enterprise rollout to tens of thousands of Accenture professionals.