OpenAI unveiled its newest family of models on Thursday, introducing a new set of heavyweights to the growing field of artificial intelligence offerings.
The GPT-5.6 comes in three variants: Sol (considered a workhorse), Terra (a more intermediate option) and Luna (a budget option). These models expand what users can do in a variety of areas—the company promises strong capabilities in enterprise work, coding, and even scientific research.
CEO Sam Altman promised that his company’s newest models are more efficient and economical than previous versions. tells CNBC that Sol AI is a 54% more efficient token when it comes to coding tasks.
Most notably, the company calls 5.6 “the most powerful cybersecurity model yet, achieving borderline performance with significantly fewer signatures.”
Indeed, there has previously been much debate about the cyber capabilities of a model like the Trump administration. was trying to limit its spreadprobably due to fear of how the model could be misused. GPT-5.6 supports defensive activities including threat modeling, code review and patching, and blue teaming (simulating an attack on your own systems to find vulnerabilities before real hackers do).
It also released a new tool called OpenAI ChatGPT caseit’s apparently designed as a workplace companion for enterprise teams across desktop, web, and mobile devices that can help with everyday clerical tasks like creating documents, spreadsheets, and presentations.
OpenAI’s newly announced family of models follows similar releases from rivals SpaceXAI and Meta this week.
However, GPT-5.6 and its accompanying marketing are most likely intended to target OpenAI’s main competitor, Anthropic. Anthropic managed to make himself one the lovable underdog AI race by focusing on enterprise customers and gaining a growing share of support as a result.
Not to be outdone, OpenAI references Artificial Analysis Coding Agent Indexa remarkable benchmark metric to claim that the latest model family outperforms Anthropic models at every step.
OpenAI calls Sol “the best coding model ever” and has openly compared it to Anthropic’s. the recently released (and much hyped) Fable. Using the Coder Agent Index, OpenAI claims that Sol “sets a new state of the art at 80, 2.8 points higher than Fable 5, while using less than half the output tokens, taking less than half the time, and spending about a third less.”
It adds: “This advantage spans the entire family: Terra outperforms Fable 5 slightly, while Luna outperforms Opus 4.8.”
The company says 5.6 is now available in ChatGPT, Codex, and the OpenAI API. Availability for one million tokens is estimated as follows: Sol $5 in / $30 out, Terra $2.50 in / $15 out, Luna $1 in / $6 out.
When you purchase through links in our articles, we may earn a small commission. This does not affect our editorial independence.
