GPT-6 Astra
GPT-6 Astra is a reasoning model from OpenAI, released in the API on September 3, 2026, which OpenAI describes as its most capable API model for complex reasoning, coding, computer use, research and document creation. It takes text and image input, returns text, and has an April 30, 2026 knowledge cutoff.
Released
September 3, 2026
Type
reasoning
License
proprietary
Context
1,050,000 tokens
Pricing
Input
$10.00 / 1M tokens
Output
$50.00 / 1M tokens
List price on the developer's own pages, checked . Standard tier, prompts up to 272K input tokens. Cached input $1, cache writes $12.50. Prompts over 272K input tokens: 2x input and cache rates and 1.5x output for the full request ($20 / $75 on the pricing page). Batch and Flex 50% of Standard; Fast mode 2x.
Capabilities
Sources
Every figure on this page was checked against these pages on . A figure we could not source is left out.
- OpenAI API docs: GPT-6 Astra (opens in a new tab)
- OpenAI API docs: Changelog (opens in a new tab)
- OpenAI API docs: Pricing (opens in a new tab)
- OpenAI API docs: Models (opens in a new tab)
Spotted something out of date? Tell us.
Links
In the news: GPT-6 Astra and OpenAI
Search Engine Roundtable
OpenAI ChatGPT Ads Account Manager Representatives
OpenAI has hired human representatives to assist advertisers with managing their ChatGPT Ads. These individuals are designated as ChatGPT Ads Account Managers. These account managers are reportedly real people who will help agencies manage ads for their respective customers on the ChatGPT Ads platform.
The Hacker News
Researchers Used a Public OpenAI Model to Build a $250,000 Chrome Full-Chain Exploit in Six Days
Security firm XBOW developed a full-chain exploit for Google Chrome by chaining three previously unknown bugs. This exploit allows code execution on a Windows computer simply by loading a web page. The researchers built the attack in six days, utilizing AI agents that were directed by OpenAI's publicly available GPT-5.6 Sol model. Google has since confirmed the existence of all three bugs, has fixed them, and has paid XBOW its top bounty. The exploit's creation highlights the potential for AI agents to accelerate the development of sophisticated cyberattack tools.
Search Engine Roundtable
OpenAI Drops UTM Parameters From ChatGPT Results (GPT-6)
OpenAI has removed or replaced UTM parameters on links within ChatGPT results, making it more difficult to track if a click originated from ChatGPT or another platform. This change appears to have begun with the release of GPT-6, which occurred approximately one week prior to this report.
OpenAI
How Third Moment Research runs 10,000 agents for 36+ hours
Third Moment Research is operating 10,000 AI agents for over 36 hours using OpenAI's GPT-6 Astra and GPT-6.1 Sol models. The AI lab employs Codex for cybersecurity tasks, enabling these agents to maintain focus for extended periods and improve their performance with use.
OpenAI
The EU’s next chapter in AI is agentic
Agentic AI adoption is growing across EU enterprises and small businesses. New analysis indicates that deeper use of agentic AI could unlock more value within these sectors.