The basicsAI glossary
GPT
"Generative Pre-trained Transformer": the name of the OpenAI model family that powers ChatGPT. Each letter sums up the method: a model that generates text, pre-trained on vast quantities of data, and built on the Transformer architecture.
In other words
The number indicates the generation: from GPT-1 in 2018 to GPT-6 today, each version was trained with far more compute than the previous one.
Alyriaz articles that mention it
- GPT-6 for everyone in ChatGPT, but not the same one
GPT-6 arrives in ChatGPT for everyone, with interactive responses. But subscribers get GPT-6 Sol, free users GPT-6 Luna. What that changes.
- GPT-6.1 Sol: the "critical" tier for 2 dollars
OpenAI ranks GPT-6.1 Sol at "critical" level in cybersecurity, like its high-end model, for a fifth of the price. What that changes.
- The model rankings as of 23 September 2026
State of the model landscape as of 23 September 2026: who leads on what, at what price, and how to read leaderboards that have become hard to compare.
- GPT-6 Sol and Luna: prices halved, the same evening as Opus 5.5
OpenAI has released GPT-6 Sol and Luna with prices halved, just hours after Claude Opus 5.5. We analyse the figures, the tests where the new model regresses, and what this convergence means.
- GPT-6 Astra: the first model rated "critical" in cybersecurity, and two versions to handle it
On 3 September, OpenAI unveiled GPT-6 Astra, the first model to reach the critical cyber capability level under its own evaluation framework. We analyse the release, the two-tier structure, the AGI claim, and a finding on reasoning monitoring.
- GPT-Live: voice without the text detour
OpenAI has launched GPT-Live, a native voice model announced with under 300 milliseconds of latency. We explain why removing the text step changes the nature of conversation, and what it raises.