OpenAI Dramatically Reduces Access Costs for ChatGPT
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
OpenAI Drastically Reduces Access Costs for ChatGPT
OpenAI has reportedly cut response costs for guest users of ChatGPT by more than half. Engineers at OpenAI informed their colleagues earlier this month that they had managed to reduce inference costs—the expense associated with running existing AI models—by over 50%. This was reported by a person familiar with the discussions, according to The Information.
OpenAI has applied these new optimizations to ChatGPT, specifically for visitors who do not have an account. The number of Nvidia GPUs required to serve these users has dropped to just a few hundred. It is unclear how many were needed previously or what techniques OpenAI used to achieve this. Guest users only have access to a very limited set of ChatGPT features, so it remains to be seen whether these gains will translate to the full product.
Deepseek has also recently launched a new open-source method that can accelerate inference requests by 60 to 85%. The freed-up resources could be used for service expansion, better models, faster responses, or larger profit margins. However, given that data center builds are progressing slowly, such gains will likely give labs more leeway rather than reduce the demand for chips.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.