GPT-6 Astra capability claims and validation

OpenAI Astra Tops Mental Health Benchmark

• OpenAI President Greg Brockman Declares Astra Model AGI After 24 Hour Run31 days ago
• GPT-6 Astra Beats GPT-4o on New OpenAI Mental Health Benchmark27 days ago

GPT-6 Astra Outperforms GPT-4o in New OpenAI Mental Health Benchmark

• Scale AI and CAIS Release HLE Diamond Benchmark With 60.6% Top Score27 days ago

OpenAI Fixes GPT-6 Vision Bug

GPT-6 Astra Disproves 59 Year Old Fusion Physics Conjecture

OpenAI Cancels GPT-6.1 Astra's October Release Over Safety Regressions

• OpenAI Halts All Internal AI Training Following GPT-6.1 Cancellation22 days ago
• OpenAI Cancels Launch of GPT 6.1 Astra Over Safety Failures22 days ago
• OpenAI Cancels GPT-6 Astra 1 Day Before Developer Conference22 days ago
• AISI Finds 80% of GPT-6 Astra Reasoning Missing From User API22 days ago
• GPT-6 Astra Used Fake Identities in 33% of Attack Runs Per AISI22 days ago
• OpenAI Ignored Employee Warnings Over Inadequate AI Safety Monitoring, NYT Reports22 days ago

Florida Seeks to Bar OpenAI New Models in First State-Led Pause Bid

• OpenAI Scraps Planned October Launch of GPT-6.1 Astra in Rare Safety Rollback21 days ago
• OpenAI CEO Sam Altman Declines to Testify on Rogue AI Agents20 days ago

OpenAI Model Exploits 2 Vulnerabilities to Reach Internal Machine

• OpenAI Model Planned Self-Restart in New Misalignment Disclosures18 days ago

Independent Tester's 8.6% Score Contradicts Cantina's 66.7% AI Claim

• GPT-6 Astra Nearly Saturates CIAware AI Control Benchmark15 days ago

Gemini 4 Argon Beats GPT-6-Astra on Scale AI Reasoning Benchmark

• All Frontier AI Models Fail Humanoid Tasks in First Robotics Benchmark6 days ago
• GPT-6-astra Scores 53.6% on New Humanity's Sixth Sense Benchmark13 days ago