据报道，OpenAI 正在研发代号为 Arrakis 的模型，该模型比 GPT-4 和 Gobi 更强大。最近，两名 Reddit 用户声称可以访问 OpenAI 的两个内部模型：Gobi 和 Arrakis，并称名为“Arrakis”的模型无所不能，超过了 GPT-4 的能力，在各个领域的表现都接近人类专家。该模型幻觉率低于 GPT-4，并且一半的训练数据是合成的，将于 2025 年发布。
OpenAI 的 GPT-4 模型已经成为人工智能领域的重要里程碑，而此次研发的 Arrakis 模型进一步扩展了 OpenAI 在自然语言处理领域的领先地位。预计，该模型的发布将对人工智能领域产生深远影响。
虽然该模型在各个领域的表现都接近人类专家，但目前仍不清楚它是否能够超越 GPT-4。不过，考虑到该模型的训练数据中有一半是合成的，其表现优于 GPT-4 的可能性是存在的。
OpenAI 表示，Arrakis 模型将于 2025 年发布，但并未透露具体发布时间。目前，该模型已经开始在内部进行测试，并表现出高度的性能。
据 Metaverse Post 报道，Arrakis 模型的训练数据中有一半是合成的，这表明该模型已经在一定程度上具备了人类专家的能力。此外，该模型的幻觉率低于 GPT-4，这也表明它能够在一定程度上避免机器生成内容的常见陷阱。
综上所述，OpenAI 研发的 Arrakis 模型比 GPT-4 和 Gobi 更强大，且在各个领域的表现都接近人类专家。预计，该模型的发布将对人工智能领域产生深远影响。
新闻翻译：
Title: OpenAI Develops New Model, Arrakis, Reportedly Stronger Than GPT-4 and Gobi
Keywords: OpenAI, new model, GPT-4, Gobi, strong
News content:
According to reports, OpenAI is developing a new model called Arrakis. This model is reportedly stronger than GPT-4 and Gobi. Recently, two Reddit users claimed to have access to OpenAI’s two internal models: Gobi and Arrakis, and并称名为 “Arrakis” model capable of doing anything, exceeding the capabilities of GPT-4 in various fields. The model’s illusiveness is lower than GPT-4, and half of the training data is synthetic, to be released in 2025.
OpenAI’s GPT-4 model has already become a significant milestone in the field of artificial intelligence, and the development of the Arrakis model further extends OpenAI’s lead in natural language processing. It is expected that the release of this model will have a profound impact on the field of artificial intelligence.
It is still unclear whether this model can surpass GPT-4 in terms of strength. However, considering that half of the model’s training data is synthetic, there is a possibility that it may outperform GPT-4.
OpenAI said that the Arrakis model will be released in 2025, but no specific time has been announced. The model has already begun internal testing and shown highly promising performance.
According to Metaverse Post, half of the training data for the Arrakis model is synthetic, suggesting that the model has already to some extent gained the ability of human experts. In addition, the model’s illusiveness is lower than GPT-4, which also suggests that it can avoid common pitfalls in machine-generated content to some extent.
Overall, the Arrakis model developed by OpenAI is stronger than GPT-4 and Gobi in terms of strength and performs close to human experts in various fields. It is expected that the release of this model will have a profound impact on the field of artificial intelligence.
新闻附加内容（英文）：
Title: OpenAI’s Arrakis Model: Stronger Than GPT-4 and Gobi?
Keywords: OpenAI, Arrakis, GPT-4, Gobi, strong
Translation:
According to recent reports, OpenAI is developing a new model called Arrakis, which is reportedly stronger than the popular GPT-4 and Gobi models. The two Reddit users claimed to have access to OpenAI’s internal models, Gobi and Arrakis,并称“Arrakis” model capable of doing anything, exceeding the capabilities of G
【来源】https://mpost.io/rumulations-of-openai-arrakis-to-be-even-more-powerful-than-gpt-4-and-gobi/
Views: 1