OpenAI’s GPT-5.6 Puts Enterprise AI on a Cost and Speed Clock
OpenAI’s new Sol, Terra and Luna models mark a shift from one flagship system to a tiered enterprise lineup, as customers push AI companies to deliver stronger performance without runaway token bills.
Topics
OpenAI on Thursday, 9 July, launched GPT-5.6, a new family of artificial intelligence models for ChatGPT, Codex and developers using its API, as the company shifts from a single flagship model to a tiered lineup built around performance, cost and speed.
The series includes Sol, OpenAI’s flagship model; Terra, a lower-cost model positioned for everyday work; and Luna, its fastest and most affordable model.
OpenAI said GPT-5.6 is designed to deliver “more intelligence from every token,” with stronger performance per dollar and better capability on demand for complex tasks.
Sol is aimed at coding, knowledge work, cybersecurity and scientific use cases. OpenAI said the model outperforms previous and competing frontier models while using fewer tokens and lowering estimated cost.
Terra is positioned as a balanced model with performance competitive with GPT-5.5, while Luna targets high-volume, cost-sensitive applications.
The launch comes as enterprise customers put more pressure on AI providers to reduce the cost of deployment.
At the Allen & Co. Sun Valley conference, OpenAI Chief Executive Sam Altman told CNBC that AI spending had become a major topic among business leaders.
“Everyone’s asking what we can do to help reduce spend or increase value,” Altman said, according to Business Insider. He said GPT-5.6 Sol is “54% more token efficient on agentic coding tasks,” though he did not specify the comparison baseline.
OpenAI said GPT-5.6 introduces a new “max” reasoning setting for deeper problem-solving and an “ultra” mode that uses multiple subagents to handle complex work across parallel tasks.
The company also said GPT-5.6 Sol sets new highs on coding and long-running professional workflow evaluations, while improving performance in cybersecurity and scientific tasks.
It said Terra and Luna outperform Anthropic’s Claude Fable 5 at about one-sixteenth the estimated cost.
The release also brings GPT-5.6 into ChatGPT Work and Codex. OpenAI said Free and Go users will get access to Terra, while Plus, Pro, Business and Enterprise users can choose among Sol, Terra and Luna.
Developers can access all three models through the OpenAI API.
OpenAI said the GPT-5.6 series includes its strongest safeguards to date, particularly for higher-risk activity, sensitive cyber requests and repeated misuse. The company said Sol does not cross its Cyber Critical threshold, though it acknowledged that benchmarks cannot capture every way a model may be used or combined with other tools.
The broader rollout followed an earlier limited preview requested by the US government. OpenAI said it had briefed US officials on the models’ capabilities and initially released GPT-5.6 to a small group of trusted partners whose participation was shared with the government.
The Guardian reported that the public release followed additional testing by the US Center for AI Standards and Innovation.
With GPT-5.6, OpenAI is giving enterprises more flexibility to choose between capability, latency and cost as competition intensifies among frontier model developers and cloud providers for corporate AI budgets.
