DeepSeek-V2.5 is the new champion among open source AI models. DeepSeek itself is an offshoot of the Chinese hedge fund High-Flyer Capital Management. The new model combines natural language processing and programming capabilities into one seemingly powerful system. According to Carl Franzen of VentureBeat, DeepSeek-V2.5 outperforms its predecessors in almost all benchmarks. The model offers …
A new open source model called OLMoE has been released by the Allen Institute for AI (AI2) in collaboration with Contextual AI. As Emilia David reports for VentureBeat, the model aims to be both powerful and inexpensive. OLMoE uses a mixed-expert architecture with 7 billion parameters, of which only 1 billion are active per input …
A new open source AI model called Reflection 70B has been introduced by Matt Shumer, co-founder of AI startup HyperWrite. As Shumer announced on the platform X (formerly Twitter), the model outperforms leading commercial systems in benchmarks. Reflection 70B is based on Meta’s Llama 3.1-70B Instruct and uses a new technique for self-correcting errors: the …
Google wants to improve the accuracy of its AI models. To avoid “hallucinations,” the company is working with partners such as Moody’s, Thomson Reuters, and ZoomInfo who will feed the AI systems with up-to-date information. A new “confidence score” is also supposed to indicate how confident the AI is in its answer being correct. With …
Meta is releasing a series of new AI models for audio, text and watermarks. Meta is also making two sizes of its Chameleon multimodal text model available for research. These models can be used to perform tasks that require both visual and textual understanding, such as image annotation.
Anthropic’s new language model Claude 3.5 Sonnet is causing a stir in the AI community. It reportedly outperforms previous models such as GPT-4 in benchmark tests and impresses users with its performance. It can handle complex tasks such as game or web development. Despite weaknesses in simple cognitive tasks, Claude 3.5 Sonnet shows the pace …
Cohere for AI releases the Aya 23 multilingual AI models with support for 23 languages and open weights. The models outperform its predecessor, Aya 101, and other open models on a variety of tasks, enabling researchers and practitioners to further develop the multilingual models and applications.
Journalists Dalvin Brown, Kara Dapena, and Joanna Stern tested ChatGPT, Claude, Copilot, Gemini, and Perplexity in everyday situations. Each chatbot was asked questions formulated by Wall Street Journal editors and columnists. The responses were evaluated by an independent panel of judges based on accuracy, usefulness and overall quality. The health category included questions about pregnancy, …
With its new Copilot+ PCs, Microsoft wants to usher in an era in which generative AI is processed directly on the device rather than in the cloud. These PCs have special hardware (Neural Processing Units, NPUs) to run AI processes locally. The first Copilot+ laptops, developed in collaboration with companies such as AMD, Intel, Qualcomm …
The Technology Innovation Institute (TII) in Abu Dhabi has released Falcon 2 11B, a new, powerful AI model that is freely available and multilingual. Falcon 2 11B outperforms comparable models such as Meta’s Llama 3.