US artificial intelligence (AI) company OpenAI announced its latest suite of AI models, is now generally available, after delaying the launch last month at the request of the US government over cybersecurity concerns.
The GPT-5.6 suite includes three models: Sol, the company's new flagship model; Terra, a balanced model for everyday tasks; and Luna, the fastest and most cost-effective model, OpenAI said in a statement.
The company says GPT-5.6 Sol sets a new standard for intelligence and efficiency in coding, knowledge-based work, cybersecurity, and science, while using fewer tokens and lowering the cost of use compared to previous leading models and competing models.
OpenAI also introduced two new capability options for the model lineup. The "max" setting gives GPT-5.6 more time to reason, check, and revise its approach, while the "ultra" setting coordinates multiple agents in parallel to handle more complex tasks.
According to the company, GPT-5.6 will be available starting Thursday on the ChatGPT platform, Codex, and OpenAI API, with the global rollout expected to continue gradually over the next 24 hours.
OpenAI said GPT-5.6 launched with "the most robust safeguards to date." The company explained that while the models have improved capabilities compared to previous models in biology and cybersecurity, they haven't yet surpassed the "critical" threshold in either category under its readiness framework.
