Meituan trains 1.6T-parameter AI model on Chinese chips, no Nvidia

Key takeaway

• Meituan has trained LongCat-2.0, a 1.6 trillion parameter AI model using only Chinese semiconductor hardware, without reliance on Nvidia chips.

• The model was trained on a cluster exceeding 50,000 domestic AI chips and processed over 35 trillion tokens.

• On some benchmarks it surpasses Western models like Gemini 3.1 Pro, though it underperforms on others, suggesting China is developing homegrown AI capabilities despite US export controls.

3 Key Points

FAQ

When was LongCat-2.0 trained and released?

The LongCat team has existed since 2023, and its first model shipped late last year. The body does not specify the exact release date of LongCat-2.0.

What hardware did Meituan use to train the model?

Meituan trained LongCat-2.0 on a cluster of more than 50,000 domestically made AI ASICs (application-specific integrated circuits). Meituan did not name the specific chip maker.

How does LongCat-2.0 compare to other AI models?

On SWE-bench Pro (59.5) and SWE-bench Multilingual (77.3), LongCat-2.0 tops Gemini 3.1 Pro and GPT-5.5 but falls short of Claude Opus 4.7 and 4.8. On IFEval (90.0), IMO-AnswerBench (81.8), and GPQA-diamond (88.9), it trails Gemini and GPT-5.5 by a wide margin in some cases.