OpenAI rolls out GPT-6 Astra to Plus, Pro, Enterprise, and Business Standard and Premium users in ChatGPT Work and Codex and in the API

OpenAI has launched GPT-6 Astra, a significant upgrade to its ChatGPT and Codex platforms, now available to Plus, Pro, Business, and Enterprise users, along with API access. This new model is positioned as the "world's most intelligent and aligned model," demonstrating state-of-the-art performance across complex tasks including computer use, software engineering, cybersecurity, and scientific research. GPT-6 Astra introduces enhanced capabilities for professional work, such as multi-step workflow execution and polished document generation, and nearly doubles the speed of computer use in ChatGPT. For Codex, a new context preservation method allows the model to retain detailed information across extended sessions, overcoming previous limitations of summary compaction. This release signifies OpenAI's continued rapid development in AI, pushing the boundaries of what is possible with large language models and offering advanced functionalities to a wider user base, particularly professionals and developers.

AI Signal Decode

GPT-6 Astra represents a substantial leap in AI capabilities, achieving exceptional scores on benchmarks like FrontierMath Tier 4 (98%) and ARC-AGI-3 (99.9%), and a perfect 100% on ExploitBench. Its advanced "computer use" feature enables ChatGPT to interact with and manage desktop environments, including Mac, at nearly twice the previous speed, facilitating demanding professional tasks. The model's proficiency in software engineering is highlighted, with OpenAI claiming it's their "best model for software engineering to date." This version also boasts a "significant jump in cyber capabilities," meeting critical preparedness thresholds, which has influenced its phased rollout for enhanced security.

The market implications are considerable, as GPT-6 Astra directly targets enhanced productivity for professionals and developers across various sectors. Its integration into ChatGPT and Codex makes these tools more powerful for complex problem-solving, coding, and cybersecurity applications. The accelerated performance and improved context handling in Codex are particularly noteworthy for software development workflows, potentially reducing development cycles. The availability to a broad range of subscription tiers, from Plus to Enterprise, suggests OpenAI's strategy to capture a larger share of the professional AI market, while API access allows for broader third-party integration and innovation.

Technically, GPT-6 Astra's advancements lie in its sophisticated handling of "computer use" and a novel context preservation mechanism for Codex. Instead of relying on summarization, Astra can maintain and search through detailed historical context, crucial for long-term projects like debugging or large-scale refactoring where retaining granular information is vital. This addresses a key limitation in LLM session management. The model's overall intelligence and alignment, built on extensive research in pre-training and reinforcement learning, position it as a formidable tool for scientific discovery and complex professional tasks, pushing the envelope on LLM reasoning and execution capabilities.

Looking ahead, the gradual rollout of GPT-6 Astra, starting with limited organizations and expanding over days, indicates a cautious approach, likely due to its advanced cybersecurity features. Users should monitor for wider availability and explore the new features, especially the improved computer use and Codex context management. The performance improvements, even on older models like GPT-5.6 (around 60% speedup), suggest ongoing optimization efforts by OpenAI. Further details on how Astra integrates into existing subscription allowances and the availability of 'GPT-6 Astra Pro' for higher tiers will be key indicators of its market penetration and value proposition.