Updated
Updated · OpenAI · Jul 9
OpenAI Launches GPT-5.6 Family at Up to 16x Lower Cost Than Fable 5
Updated
Updated · OpenAI · Jul 9

OpenAI Launches GPT-5.6 Family at Up to 16x Lower Cost Than Fable 5

3 articles · Updated · OpenAI · Jul 9

Summary

  • OpenAI began global rollout of GPT-5.6 Sol, Terra and Luna across ChatGPT, Codex and its API, with availability expanding over the next 24 hours.
  • 53.6 on Agents’ Last Exam put Sol 13.1 points above Claude Fable 5, while Terra and Luna beat Fable 5 at about one-sixteenth the estimated cost; Sol also neared Fable 5 on a broad intelligence index in 61% less time.
  • 80 on the Coding Agent Index made Sol OpenAI’s top coding model, 2.8 points above Fable 5 while using less than half the output tokens, taking less than half the time and costing about one-third less.
  • 700,000 A100e GPU hours of automated red teaming backed what OpenAI called its strongest safeguards yet; Sol’s cyber protections block roughly 10 times more potentially harmful activity than prior models.
  • $5 per 1 million input tokens prices Sol in the API, versus $2.50 for Terra and $1 for Luna, as OpenAI pushes a three-tier lineup aimed at broader enterprise and developer adoption.

Insights

What security threats prompted the government's unprecedented pre-release review of OpenAI's powerful new AI?
How will government pre-approval of AI models change the competitive landscape for smaller technology startups?
As government and AI labs collaborate on safety, what does this new approval process mean for public access?

GPT-5.6 Sol Rollout: U.S. Pre-Clearance, Cybersecurity Breakthroughs, and the Future of Frontier AI Access

Overview

OpenAI's GPT-5.6 Sol began its rollout on July 9, 2026, introducing unprecedented speed and advanced capabilities. However, broad public access is being staged, with initial availability limited to select customers as OpenAI expands capacity. This cautious approach follows a launch delay caused by heightened US government scrutiny, driven by concerns over the potential misuse of advanced AI by foreign governments or military organizations. The staged release and government involvement highlight a new era of oversight for frontier AI, balancing innovation with national security and responsible deployment.

...