Meituan and LongCat-2.0: Chinese AI Breaks Free from Nvidia
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
Meituan and LongCat-2.0: Chinese AI Breaks Free from Nvidia
Meituan has developed a 1.6 trillion parameter AI model entirely on Chinese chips, without requiring Nvidia. "LongCat-2.0 has demonstrated that we now have the capability to train large-scale models on national computing clusters," the Chinese company stated. The training was conducted on a cluster of over 50,000 locally manufactured AI ASICs and covered more than 35 trillion tokens. The LongCat team has only existed since 2023, and its first model was shipped at the end of last year.
On certain benchmarks, LongCat-2.0 outperforms leading Western models. On SWE-bench Pro (59.5) and SWE-bench Multilingual (77.3), it surpasses Gemini 3.1 Pro and GPT-5.5, but still lags behind Claude Opus 4.7 and 4.8. In other tests like IFEval (90.0), IMO-AnswerBench (81.8), and GPQA-diamond (88.9), it shows significant delays compared to Gemini and GPT-5.5 in some cases.
The message to Washington is hard to ignore. Despite the U.S. export controls in place since 2022, China seems to have produced its first competitive trillion-parameter model fully trained on domestic hardware. However, Meituan did not specify the chip manufacturer. Furthermore, the model is not yet available on HuggingFace, making independent verification difficult.
Brief IA — L'actualité IA en français
L'essentiel de l'actualité de l'intelligence artificielle, décrypté et expliqué chaque jour.