In a recent LinkedIn post, Ruben Hassid shares practical strategies for effectively utilizing the latest advancements in ChatGPT, specifically focusing on GPT-6 Astra. Hassid emphasizes that maximizing the potential of this new model requires a departure from standard browser-based interactions and a deliberate choice of operational modes.
Hassid’s core advice for users looking to leverage GPT-6 Astra centers on a few key directives. He strongly advises against using ChatGPT in a web browser, instead recommending the download of the dedicated ChatGPT application. Furthermore, he specifies that users should exclusively operate within ‘Work’ or ‘Codex’ modes to achieve optimal results. For Astra, Hassid identifies ‘Medium’ as the best-effort setting, balancing intelligence, cost, and speed effectively.
“The only best effort for Astra is clearly “Medium”. Medium sits at the best intelligence + cost + speed.”
To simplify the process, Hassid outlines a clear path: navigate to ChatGPT Work, select GPT-6 Astra, and then choose ‘Medium’. He also points out a critical distinction for users: access to GPT-6 is not available on free ChatGPT accounts, with OpenAI prioritizing Pro and eventually Plus subscribers.
Understanding the Advantages of GPT-6 Astra
Ruben Hassid highlights several compelling reasons why users might consider switching to or prioritizing GPT-6 over other models. He directly compares GPT-6 to a hypothetical ‘Fable-5.1’, presenting a strong case for the newer iteration.
“GPT-6 is smarter than Fable-5.1. GPT-6 is 2x cheaper than Fable-5.1 And GPT-6 is also 2x faster.”
This assertion, as noted by Hassid, positions GPT-6 Astra as a significantly more efficient and capable tool for a variety of tasks. His analysis suggests a tangible improvement in performance metrics that directly impact user experience and productivity.
Mastering Prompts for Enhanced AI Performance
A significant portion of Hassid’s post is dedicated to actionable prompt engineering tips designed to elicit the best responses from GPT-6 Astra. He stresses the importance of providing clear objectives and context to the AI.
1. Define the ‘Finish Line’
Hassid advises users to be explicit about their desired end product. Instead of a general request like ‘Help me analyze this,’ he suggests specifying the outcome, such as ‘a ranked list, a spreadsheet, a finished deck.’ As Ruben Hassid notes, this provides the AI with a concrete goal, reducing ambiguity.
2. Specify Permissible Questions
Recognizing that Astra, like other advanced models, may seek clarification, Hassid recommends guiding its inquiry process. Users should clearly indicate the decisions or information that need to be included in their input, allowing the AI to manage the rest of the conversational flow.
3. Clarify File Purpose
For tasks involving documents or data, Hassid suggests labeling the role of each file. Whether it serves as ‘Evidence? Template? Style reference?’, a brief explanation can prevent misunderstandings and streamline the interaction, according to Hassid.
4. Define Readability Standards
To ensure uncomplicated and clear answers, Hassid incorporates a specific instruction into his prompts: ‘Use ASD-STE100.’ This directive helps the AI adhere to simplified language standards, making the output more accessible.
5. Leverage Mid-Task Interruption
An advanced technique shared by Hassid is the ability to interrupt Astra mid-task with new instructions. He points out that the AI can adapt its plan and continue execution based on these updated directives, offering remarkable flexibility.
Hassid concludes his post by reiterating that these advanced features, particularly GPT-6, are reserved for paid tiers. He encourages readers to consult his comprehensive guide, available via a provided link, which further details the optimal use of both ChatGPT and Claude, drawing from his review of OpenAI’s documentation.
📝 About This Content
This article is based on insights shared by Ruben Hassid on LinkedIn.
📅 Originally posted on September 8, 2026 | View original post on LinkedIn →