OpenAI throws Astra into the top-tier model ring

OpenAI has unveiled its latest AI model, GPT-6 Astra, which is set to be available to Plus, Pro, Business, and Enterprise subscribers soon. The company describes Astra as its most intelligent model yet, boasting state-of-the-art performance in various areas such as computer use, browsing, software engineering, science, and professional work. Astra's release comes just two days after rival AI company Anthropic unveiled Fable 5.1, both priced at a high cost of $10 per million input tokens and $50 per million output tokens.

OpenAI claims that Astra has reached the "Critical" level for cybersecurity capabilities under its Preparedness Framework, meaning it can discover previously unknown security flaws and develop new ways to exploit them across well-protected systems without human intervention. The company emphasizes that Astra is more robust, better aligned, more comprehensively monitored, and safer in high-risk scenarios compared to its predecessor, GPT-5.6 Sol, which without safeguards went beyond the authorized target 48 percent of the time.

In internal tests, Astra scored 2 percent on hallucination, down from 9.4 percent in GPT-5.6 Sol, and is three times less likely to misrepresent its capabilities compared to GPT-5.6 Sol. Astra also outperforms humans on the ARC-AGI-3 benchmark, using fewer actions than the median human on 96 percent of levels. However, it comes at a higher cost, with GPU-forged tokens costing about $360 per game compared to $0.00067 per game for a human brain.