OpenAI GPT-5.6 Sol, Terra, and Luna Models Launched Publicly on the Same Day

Fact Restoration

On July 9, 2026, OpenAI officially launched GPT-5.6 Sol, Terra, and Luna to all ChatGPT users and API developers. These models are available on ChatGPT, the OpenAI API, and the Codex platform. The preceding 13-day government-coordinated preview, which began on June 26, concluded that day, with approximately 20 selected partner organizations participating in this preview phase. OpenAI confirmed this arrangement via the X platform on July 8, stating that GPT-5.6 Sol, along with Terra and Luna, would be publicly released on Thursday and that preview access was being expanded globally.

On the same day, SpaceXAI also publicly launched Grok 4.5, claiming it achieves Opus-level performance. Since the implementation of the Fable 5 export control ban on June 12, this marks the first time major frontier AI labs have publicly available models simultaneously. Gemini 3.5 Pro remains the only anticipated model still in preview.

Mechanism Breakdown

At the core of this release is the availability arrangement of the three models based on usage tiers. GPT-5.6 Sol is aimed at ChatGPT Pro subscribers and enterprise API users, providing access to the highest capability tier; GPT-5.6 Terra targets daily production workloads and is expected to become the default option for standard paid ChatGPT plans; GPT-5.6 Luna covers broader basic use cases. After the preview phase, the models are directly opened to global users, ending the previous restriction of limited partner access.

This arrangement, together with the simultaneous release of Grok 4.5, forms the backdrop of the July 9 event. OpenAI's announcement on the X platform clarified the timing, showing its approach of expanding access while maintaining tiered services for different user groups.

Industry Impact

In terms of competitive landscape, July 9 became the first date since the Fable 5 export control ban where all major frontier labs simultaneously have publicly available models. This situation ends the previous state where some labs' models were restricted, enabling developers to compare the latest products from OpenAI, SpaceXAI, and others at the same time.

For developers, the simultaneous availability of the three models on ChatGPT, the API, and Codex lowers the barrier to access. The tiered pricing and access settings allow developers to choose the appropriate version based on workload, with the Terra version positioned as the primary option for daily production.

For enterprise users, the high-capability access of the Sol version targets scenarios requiring peak performance, while the Luna version supports broader basic deployments. The end of the preview phase means enterprises can directly integrate the models into existing workflows without waiting for additional approvals.

Strategic Assessment (Analysis)

Based on the confirmed release arrangement, the most likely scenario going forward is that labs will further adjust their model update cadences to maintain public availability. The fact that Gemini 3.5 Pro remains in preview suggests some models still require additional validation time.

The pattern of launching three models on the same day may prompt other labs to accelerate similar tiered release strategies to cope with simultaneous competition. The developer community's response to the coexistence of multiple models will determine the direction of subsequent API usage and integration.