OpenAI has released its new family of AI models, including the flagship GPT-5.6 Sol, which CEO Sam Altman told CNBC is 54pc more token-efficient on agentic coding tasks. GPT-5.6 Sol scored 53.6 on an adaptive reasoning evaluation, beating rival Anthropic’s Claude Fable 5 by around 13 points. At medium reasoning, it beat Fable by around 11 points at roughly one-quarter of the estimated cost. OpenAI said GPT-5.6 Terra and GPT-5.6 Luna outperform Fable 5 at around one-sixteenth the cost. On the Artificial Analysis Intelligence Index, Sol at maximum reasoning takes 61pc less time at roughly half the estimated cost compared to Fable 5. The models were subject to a staggered launch at the request of the US government. OpenAI, reportedly considering giving the US government a 5pc stake in the $852bn company, expressed dissatisfaction with the government access process, stating it “keeps the best tools from users.” Altman said the company worked with US officials including commerce secretary Howard Lutnick, treasury secretary Scott Bessent and national cyber director Sean Cairncross on approval. GPT-5.6 Sol does not cross the ‘cyber critical threshold’ under OpenAI’s preparedness framework.