China's new AI model claims to beat competitors 中国新的人工智能模型宣称能击败竞争对手

DeepSeek, a Chinese artificial intelligence firm, released an open AI model last week and claimed it has beaten many foreign competitors on popular benchmarks.
中国人工智能公司 DeepSeek，在上周发布了一个开源的AI模型，并宣称在主流基准测试中已经超过很多国外的竞争对手。

The company also claimed it only spent $5.5 million to train DeepSeek V3, a fraction of the development cost of other models like OpenAI's GPT-4.
该公司还声称他们只花费了550万美元来进行 DeepSeek V3 的训练，这只是像 OpenAI GPT-4 等其他模型开发成本的一小部分。

According to DeepSeek's internal benchmark testing, DeepSeek V3 outperformed both downloadable available models and closed AI models that can only be accessed through an API.
根据 DeepSeek 内部的基准测试，DeepSeek V3 性能超过了可直接下载的可用模型，已经只能通过API访问的闭源模型。

In a subset of coding competitions hosted on Codeforces, a platform for programming contests, DeepSeek outperforms other models, including Meta's Llama 3.1 405B, OpenAI's GPT-4o, and Alibaba's Qwen 2.5 72B.
在编程竞赛平台 Codeforces 主办的一个编码竞赛子集中，DeepSeek 性能超过了包括 Meta Llama 3.1 405B，OpenAI GPT-4o，以及 Alibaba Qwen 2.5 72B 等在内的其他模型。