Latest news with #openweightmodels
Emirates 24/7
6 days ago
- Business
- Emirates 24/7
OpenAI launches two new AI models that can be customised
ChatGPT maker OpenAI has launched two new "open" artificial intelligence (AI) models that can be customised and run locally on a laptop, the company said on Tuesday. Dubbed "gpt-oss-120b" and "gpt-oss-20b," the two models are so-called "open-weight language models" which allow users to access the AI's parameters and customise them, in a departure from the company's ChatGPT bot, which is based on a "closed" model. Such models are often used by companies who use the AI tailored to their needs on their own servers. OpenAI did not say which data were used to train the new models. GPT-oss-120b can run on a graphics card, while the slimmer GPT-oss-20b version can run on a laptop with 16 gigabytes of RAM, allowing OpenAI to offer its language models for a wider range of devices. Like other text-generating AI tools, GPT-oss-120b and GPT-oss-20b are able to write code and research online when prompted. Follow Emirates 24|7 on Google News.
Geeky Gadgets
6 days ago
- Business
- Geeky Gadgets
Horizon by OpenAI : The Open-Weight Model Shaking Up the AI World
What if the future of artificial intelligence wasn't locked behind proprietary walls but instead placed in the hands of researchers, developers, and innovators worldwide? OpenAI's latest breakthrough, Horizon, promises just that—a bold step into the realm of open-weight models that could redefine how we think about accessibility and innovation in AI. With a staggering context window of 256,000 tokens and the ability to generate outputs at unprecedented scales, Horizon isn't just another large language model (LLM); it's a statement. Yet, as whispers of leaked weights and ethical dilemmas swirl, the model's release raises as many questions as it answers. Could Horizon be the key to providing widespread access to AI—or will it spark new controversies in an already heated field? Prompt Engineering explore why Horizon is more than just a technical marvel; it's a potential fantastic option for industries ranging from healthcare to content creation. You'll discover how its mixture of experts architecture and advanced coding capabilities position it as a versatile tool for solving complex problems. But the story doesn't end there. We'll also unpack the uncertainties surrounding its licensing, the implications of its open-weight nature, and what this means for the global AI landscape. Whether you're a developer eager to experiment or a researcher pondering its ethical ramifications, Horizon offers a glimpse into the future of AI—a future that's as thrilling as it is uncertain. Horizon: OpenAI's New LLM What Makes Horizon Stand Out? Horizon Beta builds upon the foundation established by Horizon Alpha, introducing several enhancements that set it apart in the competitive field of large language models. Key features include: A context window of 256,000 tokens , allowing the model to process extensive text sequences with remarkable efficiency. , allowing the model to process extensive text sequences with remarkable efficiency. A maximum output of 128,000 tokens , making it ideal for large-scale text generation tasks such as long-form content creation and detailed reports. , making it ideal for large-scale text generation tasks such as long-form content creation and detailed reports. Throughput speeds ranging from 37 to 140 tokens per second, making sure high performance for applications requiring rapid processing. These advancements position Horizon as a powerful tool for industries and tasks that demand large-scale, high-speed text generation, including content creation, data analysis, and automated documentation. Capabilities and Limitations Horizon demonstrates exceptional performance across a range of applications, making it a versatile tool for developers and researchers. Its key capabilities include: Advanced coding capabilities , such as solving complex algorithms and generating functional mock websites, which can streamline software development processes. , such as solving complex algorithms and generating functional mock websites, which can streamline software development processes. Transforming unstructured data into structured formats, a feature that is particularly valuable for industries reliant on data organization, such as finance, healthcare, and logistics. Despite these strengths, Horizon is not without its limitations. The model struggles with highly specific prompts and complex simulations, indicating areas where further refinement is needed. These challenges highlight the ongoing need for optimization to address niche use cases and improve overall reliability. Horizon OpenAI's Secret Open-Weight Model Watch this video on YouTube. Advance your skills in open source by reading more of our detailed content. Technical Insights and Speculations Horizon's architecture has sparked widespread interest and speculation within the AI community. Reports suggest the model is available in two sizes—120 billion and 20 billion parameters—providing scalability for a variety of applications. Additionally, rumors indicate that Horizon employs a 'mixture of experts' architecture with 128 experts, a design that could enhance its ability to efficiently handle diverse tasks. Adding to the intrigue, claims of leaked quantized weights on platforms like Hugging Face have fueled discussions about the model's accessibility. If these claims are accurate, they could open new opportunities for experimentation and deployment. However, such developments also raise important questions about intellectual property rights and ethical use, underscoring the need for responsible handling of open-weight models. Shifting the Open-Weight Model Landscape The release of Horizon as an open-weight model has the potential to disrupt a field currently dominated by Chinese-origin models. OpenAI's entry into this space introduces a new benchmark for performance and accessibility, fostering competition and innovation. By setting higher standards for open-weight models, Horizon could encourage broader adoption across both research and commercial sectors. This development may also accelerate advancements in AI applications worldwide. With its robust capabilities and open-weight nature, Horizon could empower researchers, developers, and organizations to explore new possibilities in natural language processing, data analysis, and beyond. The model's release signals a shift toward greater collaboration and transparency in the AI community, paving the way for future breakthroughs. Unanswered Questions and Broader Implications While Horizon's capabilities are promising, several critical questions remain unanswered. The model's licensing terms are still unclear, leaving uncertainty about whether it will be freely available for research purposes or restricted for commercial use. This ambiguity could significantly influence its adoption and impact across various sectors. Additionally, the absence of benchmark performance results raises questions about how Horizon compares to existing models in terms of efficiency, accuracy, and scalability. Without these metrics, it is difficult to fully assess the model's potential and its place in the broader AI landscape. These uncertainties highlight the importance of greater transparency from OpenAI. As Horizon gains traction, clearer details about its licensing, performance, and potential applications will be essential for researchers, developers, and organizations seeking to use its capabilities effectively. A Step Forward in AI Development Horizon represents a pivotal advancement in the evolution of open-weight models, offering new capabilities that have the potential to reshape the AI landscape. By addressing key challenges in token processing, coding, and data structuring, the model underscores OpenAI's commitment to pushing the boundaries of what large language models can achieve. As more details about Horizon emerge, its role in the broader AI ecosystem will become clearer. Whether through fostering innovation, allowing collaboration, or setting new standards for performance and accessibility, Horizon has the potential to leave a lasting impact on the field of artificial intelligence. The coming months will be critical in determining how this model is adopted and integrated into the rapidly evolving world of AI applications. Media Credit: Prompt Engineering Filed Under: AI, Technology News, Top News Latest Geeky Gadgets Deals Disclosure: Some of our articles include affiliate links. If you buy something through one of these links, Geeky Gadgets may earn an affiliate commission. Learn about our Disclosure Policy.
