OpenAI and the GPT-5.5 Mystery: A Leak Fuels Rumors
Le brief IA que les pros lisent chaque soir
Les 7 actus IA du jour, décryptées en 5 min. Gratuit.
Inclus dès l'inscription : notre sélection des meilleurs guides & comparatifs IA.
Choisis ton rythme
Gratuit · Pas de spam · Désabonnement en 1 clic
A recent leak regarding GPT-5.5 has reignited speculation around OpenAI, as the company seems to be sending mixed signals. This model, briefly spotted in Codex, has piqued the interest of developers and users alike, who are wondering if a more advanced version of ChatGPT is on the horizon.
The incident occurred when users discovered unannounced models in an internal menu of Codex, only for them to quickly disappear. This appearance took place as part of an expanded testing phase, confirmed by Rohan Varma from OpenAI, who indicated that the experimentation involved approximately 100% of Codex users. The goal of this initiative is to make the best models accessible to everyone, regardless of whether they have free or paid subscriptions.
Names such as GPT-5.5 and oai-2.1 have been spotted, fueling discussions in technical circles. OpenAI has confirmed that this extensive experimentation aims to test new models beyond the Pro subscriptions. Screenshots of these models have quickly circulated, bolstering rumors of an imminent launch.
Signals from Sam Altman
Sam Altman, CEO of OpenAI, has not sought to quell speculation. He even responded to messages on social media mentioning a possible launch of GPT-5.5, without providing any clarification. One user mentioned a potential release on Thursday, to which Altman replied with a simple emoji, leaving room for interpretation.
Developers who have had a glimpse of these models speak of notable improvements, including increased speed and more efficient management of front-end tasks. Technologies like HTML and Tailwind CSS seem to benefit from these advancements, yielding cleaner and faster results.
At the same time, OpenAI continues to make product announcements, such as the ChatGPT Images 2.0 update, which enhances the accuracy of visuals generated by the tool. This momentum helps maintain constant media pressure around the company.
This leak comes as competition in the AI field intensifies, with players like Anthropic adjusting their offerings, such as limiting access to Claude Code for certain Pro users. Each company is doubling down on efforts to improve its technologies and capture user attention.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.