Mistral Small 3 的发布同时，非营利 AI 机构 A2I 也发布了新的大语言模型。Tülu 3 405B 是 Meta 去年 6 月推出的开源模型 Llama 3.1 405B 的定制版本。在 Ai2 进行的测试中，Tülu 3 405B 在超过半打基准测试中的表现都优于原始 Llama 模型。

今日（2月4日）凌晨，Chatbot Arena LLM Leaderboard更新了最新一期的榜单，不久前发布的Qwen2.5-Max直接冲进前十，超越DeepSeek V3, ...

2月4日凌晨，三方基准测试平台Chatbot Arena公布了最新的大模型盲测榜单，一周前刚发布的Qwen2.5-Max超越DeepSeek V3、o1-mini和Claude-3.5-Sonnet等模型，以1332分位列全球第七名，也是非推理类的中国大模型冠军。同时，Qwen2.5-Max在数学和编程等单项能力上排名第一，在硬提示（Hard prompts）方面排名第二。 据了解，Chatbot ...

As the battle between OpenAI and DeepSeek escalates, some Indian start-ups have started integrating DeepSeek with offerings.

New data lifts the lid on private tech investment in Australia, with a quarter of the money going to firms building products ...

Financial writer highlights META's impressive growth, AI investments, and revenue increase, projecting bullish momentum for ...

TOPS (trillion operations per second) or higher of AI performance is widely regarded as the benchmark for seamlessly running ...

Those looking for an AI tool to facilitate illegal activities might want to turn to DeepSeek. Experts have shown that the ...

Mistral Small 3 combines efficiency, adaptability, and cost-effectiveness, making it a game-changer in AI development and applications.

Paid subscribers to ChatGPT nearly tripled to 15.5 million last year from 5.8 million a year earlier, OpenAI recently told ...

Since its worldwide release last week, accounting AI experts have been testing Chinese LLM DeepSeek on accounting questions ...