China’s Meituan introduces LongCat-2.0, a large language model it says is built to run without Nvidia chips. NDTV reports that Meituan claims the model has 1.6 trillion parameters and that it performs well on coding and “agentic” benchmarks, surpassing Google’s older Gemini 3.1 Pro on multiple tests. However, NDTV adds that LongCat-2.0 remains behind some global frontier models, including OpenAI’s GPT-5.5 and Anthropic’s Claude 4.8 Opus. Channel NewsAsia (CNA) similarly reports Meituan’s claim that LongCat-2.0’s performance is comparable to Gemini 3.1 Pro, referencing Gemini’s February release. Both outlets frame the announcement as part of Meituan’s push to develop advanced AI capabilities using domestic or non-Nvidia hardware. The reports focus on Meituan’s performance comparisons and do not detail independent verification, training methods beyond the chip claim, or how the benchmarks were selected.
Meituan unveils LongCat-2.0 AI model trained without Nvidia chips
China’s Meituan introduces LongCat-2.0, a large language model it says is built to run without Nvidia chips. NDTV reports that Meituan claims the model has 1.6 trillion parameters and that it performs...
- Meituan announces its LongCat-2.0 large language model.
- Meituan says the model is trained without Nvidia chips, using domestic chips.
- Meituan claims LongCat-2.0 performs on coding and agentic benchmarks.
- Meituan compares LongCat-2.0’s results to Google’s Gemini 3.1 Pro, released in February.
- NDTV adds that Meituan’s model is claimed to trail some global frontier models such as OpenAI’s GPT-5.5 and Anthropic’s Claude 4.8 Opus.
Meituan said LongCat-2.0 outperformed Google's older Gemini 3.1 Pro on several coding and agentic benchmarks, though it trailed global frontier models including OpenAI's GPT-5.5 and Anthropic's Claude 4.8 Opus.
5 hours agoMeituan says the performance of its new large language model, LongCat-2.0, is comparable to Google's Gemini 3.1 Pro, which was released in February.
1 day ago
Aflac discloses data breach after attackers compromise Japan subsidiary systems
Aflac discloses a data breach that affects customers of its Japan unit after attackers compromise systems belonging to a...
14-year-old Jainam Jain founds AI startup and works from Dubai’s Burj Khalifa
Multiple outlets profile 14-year-old Jainam Jain, who has been described as Dubai’s youngest AI startup founder. The rep...
Anthropic restores global access to Claude Fable 5 and Mythos 5 after U.S. lifts export controls
Anthropic says it is restoring global access to its Claude Fable 5 and Claude Mythos 5 models after the U.S. Department...