Skip to main content

ChatGPT & OpenAI News

Stay updated with the latest ChatGPT features, OpenAI announcements, and AI industry developments.

Latest Updates​

🎉 ChatGPT Atlas Browser (Oct 21, 2025)​

OpenAI introduced ChatGPT Atlas, a new browser with ChatGPT built directly into the browsing experience. This marks a significant shift in how users interact with AI while browsing the web.

Key Features:

  • Native ChatGPT integration in browser
  • Seamless AI-assisted web navigation
  • Enhanced productivity workflows
  • Privacy-focused design

Learn more about ChatGPT Atlas →

🎬 Sora 2 Released (Sep 30, 2025)​

The next generation of OpenAI's video generation model is here. Sora 2 brings significant improvements in video quality, length, and creative control.

What's New:

  • Higher resolution video output
  • Extended video length (up to 60 seconds)
  • Better motion consistency
  • Enhanced prompt understanding
  • Improved character consistency

Availability: Currently available to ChatGPT Plus and Pro users.

Read the Sora 2 research blog →

⚡ gpt-realtime & Realtime API Updates (Aug 28, 2025)​

OpenAI introduced gpt-realtime, a new model optimized for real-time voice and conversational AI applications.

Key Features:

  • Ultra-low latency responses
  • Natural conversation flow
  • Improved voice quality
  • Better context handling
  • Support for interruptions and turn-taking

API Updates:

  • Streaming improvements
  • WebSocket support
  • Enhanced audio processing
  • Lower pricing for real-time use cases

Use Cases:

  • Voice assistants
  • Customer service bots
  • Interactive tutoring
  • Live translation
  • Gaming NPCs

Explore the Realtime API docs →

🚀 GPT-5 Launched (Aug 7, 2025)​

The most significant model update from OpenAI. GPT-5 represents a major leap in AI capabilities.

Major Improvements:

  • 2x faster than GPT-4
  • 50% better reasoning on complex tasks
  • Extended context window: 1M tokens (API)
  • Multimodal: Text, images, audio natively supported
  • Better instruction following
  • Reduced hallucinations

Pricing:

  • Input: $3.00 per 1M tokens
  • Output: $15.00 per 1M tokens
  • Same price as GPT-4, better performance

Model Variants:

  • gpt-5 - Standard model
  • gpt-5-thinking - Enhanced reasoning
  • gpt-5-pro - Maximum capabilities (Pro tier only)

Read the GPT-5 announcement →

Recent Product Updates​

Codex Generally Available (Oct 6, 2025)​

OpenAI's Codex agent is now generally available to all ChatGPT Plus, Pro, and Business users.

Capabilities:

  • Write and debug code
  • Explain complex codebases
  • Refactor and optimize code
  • Generate unit tests
  • Terminal command assistance

Supported Languages: Python, JavaScript, TypeScript, Java, C++, Go, Rust, and 20+ more.

Try Codex in ChatGPT →

Apps in ChatGPT & Apps SDK (Oct 6, 2025)​

Developers can now build native apps that run inside ChatGPT using the new Apps SDK.

What You Can Build:

  • Custom calculators
  • Data visualization tools
  • Games and interactive experiences
  • Specialized productivity tools
  • Domain-specific assistants

Framework: React-based, deployed to ChatGPT directly.

Read the Apps SDK docs →

Research & Safety​

Political Bias in LLMs Study (Oct 9, 2025)​

OpenAI published research on defining and evaluating political bias in large language models.

Key Findings:

  • Models can reflect biases from training data
  • New evaluation framework introduced
  • Mitigation techniques tested
  • Transparency improvements

Goal: Build more neutral, balanced AI systems.

Read the research paper →

Expert Council on Well-Being and AI (Oct 14, 2025)​

OpenAI launched an Expert Council focused on AI's impact on human well-being.

Focus Areas:

  • Mental health implications
  • Social connection
  • Productivity and work-life balance
  • Educational outcomes
  • Accessibility

Members: Leading psychologists, educators, and technologists.

Company News​

OpenAI-Broadcom Partnership (Oct 13, 2025)​

Strategic collaboration to deploy 10 gigawatts of OpenAI-designed AI accelerators.

Impact:

  • Faster model training
  • Lower inference costs
  • Better availability
  • Custom silicon for AI workloads

OpenAI-AMD Partnership (Oct 6, 2025)​

Partnership to deploy 6 gigawatts of AMD GPUs for AI infrastructure.

Benefits:

  • Increased compute capacity
  • Geographic redundancy
  • Cost optimization
  • Performance improvements

Japan Economic Blueprint (Oct 22, 2025)​

OpenAI released a comprehensive plan for AI adoption in Japan.

Initiatives:

  • Japanese language optimization
  • Local data centers
  • Educational programs
  • Business partnerships
  • Research collaborations

Earlier Major Releases​

GPT-4.5 (Feb 27, 2025)​

Interim update between GPT-4 and GPT-5.

Improvements:

  • Better reasoning on STEM topics
  • Enhanced code generation
  • Improved multilingual support
  • Faster response times

OpenAI o3-mini (Jan 31, 2025)​

Lightweight reasoning model for cost-effective applications.

Features:

  • Fast inference
  • Low cost ($1/1M input tokens)
  • Good for simple tasks
  • API-only model

Model Spec v2 (Feb 12, 2025)​

Updated guidelines for model behavior.

Changes:

  • Clearer safety boundaries
  • Better response formats
  • Improved refusal handling
  • Enhanced transparency

How to Stay Updated​

Official Channels​

📧 OpenAI Newsletter: Subscribe here 🐦 Twitter: @OpenAI 📺 YouTube: OpenAI Channel 📝 Blog: openai.com/news

Community Resources​

💬 Community Forum: community.openai.com 📚 Developer Docs: platform.openai.com/docs 📖 OpenAI Cookbook: cookbook.openai.com 🔬 Research Papers: openai.com/research

Subscribe to ChatGPTLog​

Get curated updates, analysis, and practical guides delivered to your inbox.

What You'll Get:

  • Weekly digest of ChatGPT news
  • Exclusive tips and tutorials
  • Early access to new guides
  • Community insights

Subscribe to our newsletter →

Last Updated: October 23, 2025

Information sourced from official OpenAI announcements and verified news sources.