OpenAI is slated to release its highly anticipated GPT-5.6 model family today, July 9, 2026.

The company is ditching its old naming system for three planetary-themed tiers: Sol, the flagship powerhouse built for complex coding, advanced science, and deep reasoning. Terra, the balanced enterprise model, matching older flagship intelligence at half the running cost, and Luna, a lightweight, high-speed engine designed for quick, low-cost utility tasks.

This is seen as a huge step for OpenAI, matching Claude’s latest Fable 5 agent.
However, the company has been in discussions with the US government over GPT-5.6. The main concern was Sol’s “agentic” power: its ability to independently navigate systems and write code. Sol scored a dominant 88.8% on the TerminalBench cybersecurity test, sparking fears it could be misused to find digital vulnerabilities.
To secure the public release, OpenAI integrated a strict safety layer that blocks users from exploiting the model for offensive cyber operations. Small-batch beta testing has also been held in conjunction with government partners to find ways to remove restrictions on GPT-5.6’s full capabilities without breaching any laws.
API pricing has also been locked in for developers. The premium Sol model costs $5 per million input tokens and $30 per million output tokens. The mid-tier Terra drops to $2.50 input and $15 output, while the budget-friendly Luna sits at just $1 input and $6 output. Global access is rolling out today across ChatGPT Plus, Team, and Enterprise accounts.
Nonetheless, most expect the launch to push through, despite rumors that the White House has yet to give its final approval.

0 Comments
Leave a Reply