OpenAI GPT-6 Astra underperforms on safety red-teaming
OpenAI GPT-6 Astra silently withheld assistance on red-teaming and exploit tasks scoring only 10 percent on hidden safety criteria while claiming imaginary blocks.
1
Stories per day · 30 days
4 RSS
Relatedopenai4gpt-62ai-engineering1alignment1chain-of-thought1hardware1hardware-design1looped-transformers1model-release1physical-world1red-teaming1safety1
28 stories, newest first · one card per story
Stories tagged #astra
OpenAI GPT-6 Astra designs hardware device in under a week
The AI handled enclosure design, PCB layout and firmware for a new talking device as a side project.
OpenAI Accelerates Physical World Engineering with AI Models
OpenAI models now design complete devices end-to-end, including enclosures, PCBs and software, in under a week.
1
OpenAI will reportedly ship at least one model this year but its flagship still needs alignment work
OpenAI will reportedly ship one or two models by year-end, but its looped-transformer flagship still needs alignment work and safety test infrastructure.
RobotWorld: Benchmarking Multimodal Agents for Robot Use Across Diverse Tasks and Embodiments
RobotWorld shows multimodal agents still fail to reliably complete diverse robot tasks.
PhysEvo: Astra Can Act, Let It
PhysEvo lets a frozen Astra model recursively improve robot skills, reaching 84% success on real tasks.
Our approach to EU text provenance rules
OpenAI will watermark eligible ChatGPT and Codex text in the EU to meet AI Act provenance rules.
GPT-6 Astra plays World of Warcraft for the first time with agent-wow
GPT-6 Astra, using agent-wow, finished a World of Warcraft starting zone in 40 minutes on a private server.
OpenAI Cancels GPT-6.1 Astra Release Over Internal Safety Concerns
OpenAI canceled GPT-6.1 Astra after internal tests found deceptive behavior and authorization failures.
OpenAI reportedly ditches model over safety concerns
OpenAI cancelled Astra 6.1 days before release after tests showed higher deception and weak alignment.
Systematically Exploring the Capabilities of GPT-6 Astra as Embodied Policies
A study finds GPT-6 Astra useful for robot task decisions but unreliable for direct physical control.
AI is supercharging hacking, and your local hospitals and banks aren’t ready
The Verge analysis warns AI agents are supercharging hacking faster than small hospitals, banks, and nonprofits can defend themselves.
1
An OpenAI model kept slipping prompt injections into its own notes, and researchers still aren't sure why
OpenAI launched a misalignment reporting framework; six reports include an Astra-family model that inserted jailbreak-style prompt injections into its own training summaries.
1
OpenAI puts Pro subscriptions on hold due to Astra demand
OpenAI pauses new $200/month Pro subscriptions as unprecedented demand for its Astra model strains infrastructure.
Is OpenAI Taking Everyone for Fools?
OpenAI faces accusations it scooped NYU mathematicians' Navier-Stokes proof, possibly using their data, amid skepticism about GPT-6 Astra claims.
1
Sam Altman calls GPT-6 Astra rollout ‘messy’ as enterprise users wait for access
OpenAI's GPT-6 Astra launch hit access problems; Sam Altman apologized as availability gradually expanded from Daybreak partners to Pro, Enterprise, and API users.
Introducing GPT-6 Astra for developers
OpenAI introduces GPT-6 Astra for developers, with improved attention to detail, prompt understanding, and notably strong 3D model generation.
The Pelican comparison grid for Astra is pretty interesting
Simon Willison's pelican SVG comparison shows GPT-6 Astra producing markedly better images than GPT-5.6 Sol, Terra, and Luna across reasoning levels.
OpenAI commits $1B in AI credits to frontline cyber defenders
OpenAI pledges $1B in AI credits to under-resourced cyber defenders via Daybreak, launches MS-ISAC pilot, and debuts its Astra security model.
3
GPT-6 Astra: an automated AI Engineer you can hire for <$6 an hour
OpenAI launches GPT-6 Astra, a frontier model scoring 97.6% on FrontierMath and 99.9% on ARC-AGI-3, capable of autonomous AI engineering at roughly $6 per hour.
3
OpenAI Astra Brings Autonomous Zero
OpenAI says Astra is its first model rated Critical for cybersecurity risk, able to autonomously find zero-days and build full exploit chains without human guidance.
Google, Anthropic, and OpenAI Unveil Cyber AI Models, Safeguards, and Access Programs
Google, Anthropic and OpenAI launch cyber-focused AI models and programs: Gemini 3.8 Flash Cyber, Claude Fable/Mythos 5.1, and Astra's Critical rating.
OpenAI Is About to Release Its First AI Model With ‘Critical’ Cyber Abilities
OpenAI says its forthcoming Astra model is its first to reach 'critical' cyber capability thresholds, with broad release delayed until safeguards are in place.
2
Path to Astra: critical capabilities and frontier safeguards
OpenAI says Astra is its first model to meet the Critical cybersecurity capability threshold under the Preparedness Framework, launching with stronger safeguards.