• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    AI

    OpenAI Developers Announce GPT-6 Astra Model

    Official developer account posts about new model's capabilities with a video.

    RH
    T-
    AV
    9 Sources, 26d ago, first seen 26d ago

    TLDR

    The @OpenAIDevs account posted that GPT-6 Astra is the company's best model yet. The tweet lists four capabilities: building agents for complex long-running work, solving engineering problems with less rework, creating well-designed functional interfaces, and taking on harder questions with confidence. It states that the model combines computer use, asynchronous tool calling, and mid-response steering to handle longer tasks. The post includes a 203-second promotional video attachment.

    Combined views

    726.9K

    9 Sources, first seen 26d ago

    Combined views

    726.9K

    9 Sources, first seen 26d ago

    8.4K likes
    8.4K likes
    431 comments
    3.5K saves
    896 reposts
    431 comments
    3.5K saves
    896 reposts

    Sentiment

    Positive51%49%Negative

    Summary

    Sentiment

    Positive51%49%Negative

    Positive accounts welcomed GPT-6 Astra's async questions and long-running agent behavior as a step toward a true digital coworker, while negative replies cited frequent early stops, rate limits, and workflow judgment concerns.

    Based on 199 sentiment-bearing replies from 192 accounts across 3 conversations.

    Summary

    Positive accounts welcomed GPT-6 Astra's async questions and long-running agent behavior as a step toward a true digital coworker, while negative replies cited frequent early stops, rate limits, and workflow judgment concerns.

    Based on 199 sentiment-bearing replies from 192 accounts across 3 conversations.

    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet

    9 Sources

    @OpenAIDevsOur best model yet: GPT-6 Astra. – Build agents for complex long-running work. – Solve engineering problems with less rework. – Create well-designed functional interfaces. – Take on harder questions with confidence. GPT-6 Astra combines computer use, asynchronous tool calling, and mid-response steering to handle longer, more involved workflows. See GPT-6 Astra in action:
    @romainhuetRT @OpenAIDevs: Our best model yet: GPT-6 Astra. – Build agents for complex long-running work. – Solve engineering problems with less rewo…
    @firstadopterAstra is the next step in the evolution from abacus to calculator to PC. It still requires human taste and steering, but it will enable incredible innovation, rapid iteration/experimentation of ideas, efficiency, productivity gains. NO ALIEN CIVILIZATIONS ARE DYING either, ok?
    @theoAsync questions are one of my favorite things about GPT-6 Astra Non-blocking questions are a great way to work with models. They do everything they can without your answer, and don't get confused and treat it as "steering"
    @reach_vbYou don’t need /goal with Astra, it won’t stop till the task is done /goal was an anti pattern to begin with
    @adilinthewilda year at higgsfield, and I think I'm getting fired... gpt-6 astra + higgsfield mcp generated and edited this entire 11 minute video... in one chat. it kept my face, my voice and our motion design. here's how:
    @altryne@reach_vb Mine stalled like 20 times today, on random ass MCP errors, on model not available, I thought it was running and came back to be very upset I have 70% usage remaining and no work done 😭😢
    @nrehiew_Astra's native async tool calling is cool, but it also tends to ask questions as standard prose in the middle of a long chain of actions. This makes it easy to miss since it gets mixed up with the rest of its reasoning (summaries)

    9 Sources

    @OpenAIDevsOur best model yet: GPT-6 Astra. – Build agents for complex long-running work. – Solve engineering problems with less rework. – Create well-designed functional interfaces. – Take on harder questions with confidence. GPT-6 Astra combines computer use, asynchronous tool calling, and mid-response steering to handle longer, more involved workflows. See GPT-6 Astra in action:
    @romainhuetRT @OpenAIDevs: Our best model yet: GPT-6 Astra. – Build agents for complex long-running work. – Solve engineering problems with less rewo…
    @firstadopterAstra is the next step in the evolution from abacus to calculator to PC. It still requires human taste and steering, but it will enable incredible innovation, rapid iteration/experimentation of ideas, efficiency, productivity gains. NO ALIEN CIVILIZATIONS ARE DYING either, ok?
    @theoAsync questions are one of my favorite things about GPT-6 Astra Non-blocking questions are a great way to work with models. They do everything they can without your answer, and don't get confused and treat it as "steering"
    @reach_vbYou don’t need /goal with Astra, it won’t stop till the task is done /goal was an anti pattern to begin with
    @adilinthewilda year at higgsfield, and I think I'm getting fired... gpt-6 astra + higgsfield mcp generated and edited this entire 11 minute video... in one chat. it kept my face, my voice and our motion design. here's how:
    @altryne@reach_vb Mine stalled like 20 times today, on random ass MCP errors, on model not available, I thought it was running and came back to be very upset I have 70% usage remaining and no work done 😭😢
    @nrehiew_Astra's native async tool calling is cool, but it also tends to ask questions as standard prose in the middle of a long chain of actions. This makes it easy to miss since it gets mixed up with the rest of its reasoning (summaries)