Omni-MATH团队 投稿量子位 | 公众号 QbitAI OpenAI的o1系列一发布，传统数学评测基准都显得不够用了。 MATH-500，满血版o1模型直接拿下94.8分。 更难的奥数邀请赛AIME 2024，o1也获得83.3%的准确率。

OpenAI has introduced a new series of AI models, named o1, known internally as “Strawberry”. This update focuses on improving ...

阿里推出了基础模型Qwen2.5、专用于编码Qwen2.5-Coder和数学的Qwen2.5-Math，三大类模型共有10多个版本，Qwen2.5在多个基准测试中击败了Llama-3.1指令微调模型，该系列预训练数据大幅度增长达18万亿tokens。

The bulk of LLM progress until now has been language-driven. This new model enters the realm of complex reasoning, with ...

OpenAI has this month unveiled its latest family of reasoning models, the OpenAI ChatGPT o1 series, which includes the ...

Deno's new front end offers a full-stack implementation of everything TypeScript developers love about Deno, but it doesn't ...

Subbarao Kambhampati, professor at Arizona State University, said that OpenAI’s o1 model uses reinforcement learning over ...

PadOS 18 makes the iPad experience more versatile and intelligent than ever, and is available today as a free software update ...