OpenAI launches Astra, its powerful (and controversial) new model
OpenAI released Astra on Thursday, its latest AI model and — according to the company — its most powerful and capable one yet. OpenAI claims that Astra represents “a new frontier on computer and browser use,” and that it handles tasks with unmatched “speed, accuracy, and safety.” The model is being made available Thursday to…
Opening of the original on TechCrunch AI
Summary
OpenAI launched Astra, its most powerful AI model yet, claiming unmatched speed, accuracy, and safety. Available to cybersecurity and paid customers, Astra excels in identifying zero-day exploits and coding tasks, outperforming competitors like Anthropic's Fable. However, its use of opaque recurrence, a reasoning technique that obscures monitoring, makes it controversial. OpenAI president Greg Brockman suggested the model's capabilities might represent a personal threshold for AGI.
Why it matters
Astra's release signals OpenAI's push for advanced AI capabilities, particularly in cybersecurity and software engineering. Its claimed superiority in identifying zero-day exploits and coding benchmarks challenges rivals like Anthropic. The controversy surrounding opaque recurrence highlights a growing tension between model power and transparency, a key concern for researchers and users. OpenAI's framing of AGI as a "mission concept" rather than a strict definition also reflects the evolving landscape. Future developments will focus on Astra's real-world impact and OpenAI's ability to address safety concerns.
Related: Google: Introducing Gemini 3.8 Flash and 3.8 Flash Cyber · Anthropic: GPT‑6 Astra · Microsoft: Claude Fable 5.1 is generally available in GitHub Copilot · Meta: Meta Muse Code & Muse Spark Course – Build AI Agents, APIs, and Full-Stack Apps · xAI: Sam Altman "AGI by December"
Rated high: a launch, deal or policy change that changes what you can do or what it costs.
- OpenAI begins rolling out Astra model after warning of its advanced cyber capabilitiesCNBC Technology · Web · Sep 3, 2026
- OpenAI’s new reasoning technique alarms AI safety expertsTechCrunch AI · Web · Sep 2, 2026
- Researchers fear safety disaster ahead of OpenAI’s Astra releaseThe Verge AI · Web · Sep 2, 2026
- OpenAI says Astra AI model is its first that crosses 'Critical' cybersecurity capabilityCNBC Technology · Web · Sep 1, 2026
- OpenAI hails ‘new era of artificial general intelligence’ with Astra model releaseThe Guardian AI · Web · Sep 3, 2026
- GPT-6 Astra Is Here—and OpenAI Thinks It May Kick Off the AGI EraWIRED AI · Web · Sep 3, 2026
- OpenAI’s new reasoning technique alarms AI safety expertsTechCrunch AI
- Researchers fear safety disaster ahead of OpenAI’s Astra releaseThe Verge AI
- GPT-6 Astra Just Went CRITICAL...Wes Roth
- Fable 5.1 just smoked ASTRA...Wes Roth
- OpenAI says Astra AI model is its first that crosses 'Critical' cybersecurity capabilityCNBC Technology
- Path to Astra: critical capabilities and frontier safeguardsOpenAI News
Questions people ask
- What makes Astra controversial?
- Astra uses opaque recurrence, a reasoning technique that obscures the "chain of thought" process, making it harder for researchers to audit how the model reaches its decisions.
- How does Astra compare to other models?
- OpenAI claims Astra is the best model for software engineering and scores higher than models like OpenAI's Sol and Anthropic's Fable on cybersecurity benchmarks.
- Is Astra considered AGI?
- OpenAI president Greg Brockman stated that the contractual definition of AGI is no longer relevant, but personally believes Astra's capabilities might meet the definition of AGI.
More from TechCrunch AI 2 more
Page generated Sep 3, 2026. Summaries are Subvolts' own; the story belongs to TechCrunch AI.