OpenAI Launches GPT‑6 Sol and Luna with Up to 50% Lower API Prices
What Happened — OpenAI released two new GPT‑6 variants, Sol and Luna, across ChatGPT Work, Codex, and the desktop app. Both models claim superior performance on high‑effort tasks while cutting API costs by roughly half compared with prior generations.
Why It Matters for Trust & Control Assurance
- Cost‑efficient AI models change the economics of large‑scale automation, prompting organizations to revisit AI‑model selection controls and budgeting processes.
- Continuous control‑assurance programs must capture model performance, cost, and usage metrics to demonstrate due diligence and maintain a defensible audit trail.
- Mapping the new models to existing AI‑governance objectives (e.g., model risk, data provenance) is essential for compliance with frameworks such as NIST AI RMF or ISO 42001.
Who Is Affected — Technology‑SaaS vendors, enterprise AI teams, fintech platforms, and any organization that integrates OpenAI APIs into production workloads.
Recommended Actions
- Update your AI‑model procurement policy to include cost‑efficiency and performance benchmarks.
- Enable logging of token usage, model version, and task classification for continuous monitoring.
- Map the new Sol/Luna controls to your framework of record and capture evidence in your Trust Center. Source: https://www.helpnetsecurity.com/2026/09/23/gpt-6-sol-luna-lower-api-prices/
Technical Notes — The models are accessed via the gpt-6-sol and gpt-6-luna endpoints. OpenAI reports up to 91 % lower cost per task on AutomationBench and a 58 % reduction on high‑effort workloads. No new CVEs or vulnerabilities are disclosed. Source: https://www.helpnetsecurity.com/2026/09/23/gpt-6-sol-luna-lower-api-prices/