GPT-4o

For ToTheMoon, GPT-4o is the model by which the channel measures ChatGPT's path from a striking announcement to an everyday tool. It appears as the product that made AI noticeably easier for ordinary people, as the default model in arguments about which version to pick for a task, and as the model users turned out to be emotionally attached to. The hosts' question is what in GPT-4o was actually valuable, the raw capability or the way it talked to people.

The first analysis is episode 007, in May 2024. The hosts argued that GPT-4o proved the mode of interaction matters as much as power: voice, images and natural dialogue bring a model closer to a daily assistant. They also noted that nobody polices model quality, which sometimes drops. By 2025 GPT-4o had become the baseline. In episode 049 one host held that it was already good enough for most tasks, and episode 061 described it as the everyday option with o3 reserved for important work, faulting ChatGPT for making users pick the mode by hand.

The later mentions are about attachment and succession. In episode 067 Voice Mode with GPT-4o was called a pain, but in episode 069 a medical case built on the previous year's GPT-4o delivered strong results, which the hosts read as headroom for the technology. Episode 071, after the GPT-5 launch, covered users complaining about losing the more emotionally engaged GPT-4o and the model being restored at least for paying subscribers. By 2026, in episode 098, GPT-4o is a benchmark of the past: the hosts suggested comparing what one could do with it against what current models can do.