Bytedance Develops a Ten Trillion AI Model
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
Bytedance Develops a Ten Trillion Parameter AI Model
Bytedance is training an AI model that could reach ten trillion parameters, according to the Financial Times. This represents three times the size of Moonshot's Kimi K3, currently the largest Chinese model. This would place TikTok's parent company in the same category as Anthropic's Mythos 5 system, which industry estimates suggest is around eight trillion parameters. Anthropic has not disclosed its own figures.
Three internal sources told the FT that the model is in the pre-training phase, a step that typically takes between three and six months. The parameters determine the amount of information a model can store, but performance also depends on the quality of the data and training methods. One source mentioned that Bytedance has avoided distillation, meaning training on the outputs of other companies' models, for over a year.
Founder Zhang Yiming has tasked the 2,000-person Seed team with aiming for long-term world-class model capabilities. xAI is also training variants of Grok with six and ten trillion parameters on its Colossus 2 cluster, according to Elon Musk.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.