OpenAI Releases GPT-6 Astra Model
The new model introduces parallel agents and achieves faster performance across multiple computing benchmarks.
Updated on Oct. 26, 2026 in Artificial Intelligence
Live Poll
Do you trust the development of AI agents capable of performing human-level work at expert speeds?
On September 3, 2026, OpenAI launched its GPT-6 Astra model, which operates software by navigating interfaces to complete multi-step tasks. The update represents a shift toward more efficient AI, featuring a parallel agent release in mid-October 2026.
Why it matters
The model demonstrates significant gains in speed and accuracy compared to its predecessor, GPT-5.6 Sol, while utilizing fewer tokens for complex reasoning tasks. This development signals a move toward AI systems capable of executing simultaneous workflows.
Astra achieved a 92.7% score on the ScreenSpot-Pro benchmark compared to 76.9% for its predecessor, while completing tasks in 40 minutes versus 75 minutes. Additionally, the model scored 99.9% on the ARC-AGI-3 reasoning benchmark.
The players
OpenAI
An artificial intelligence research organization that develops advanced machine learning models and large language platforms.
The details
Astra functions by clicking through user interfaces and moving between applications to execute complex jobs autonomously. The mid-October 2026 update further enhanced these capabilities by allowing multiple AI workers to tackle different parts of a job simultaneously.
Timeline
September 3, 2026: The GPT-6 Astra model was officially released.
Mid-October 2026: OpenAI deployed an update featuring parallel agents.
The Tech Race
The development of Astra reflects a broader industry shift toward AI agents that can manage entire software workflows rather than just generating text. This moves the technology away from simple language processing and toward autonomous digital labor.
Users can expect faster multi-step task completion and the ability to run multiple AI processes simultaneously for complex workflows. These capabilities may reduce the time required to manage software applications through improved automated navigation.
The takeaway
The transition to parallel agents allows users to delegate multiple distinct tasks to AI systems at once, increasing efficiency. Readers should consider identifying repetitive interface-based tasks that could be automated by these new agent capabilities.
Further reading
Learn more about advancements in Artificial Intelligence.
Live Poll
Do you trust the development of AI agents capable of performing human-level work at expert speeds?
