FlagThis AI

OpenAI’s new o1 model and its Pro version offer enhanced math, coding and image processing capabilities. The o1 model is available to ChatGPT Plus and Team users, whereas the Pro version offers more advanced features. This upgrade marks a significant step in AI model evolution, showcasing improved reasoning and multimodal functionality. The Pro version of the model appears to use multiple attempts to get better answers, and offers significantly increased usage, higher resolution, and longer duration options.

OpenAI has launched an upgraded ‘o1’ reasoning model, which is currently available to top developers via their API. The new model comes with enhanced capabilities, including function calling, JSON structured output, and image analysis capabilities. This upgrade signifies continued advancements in AI technology, making more sophisticated tools accessible to a select group of developers. The rollout includes fine-tuning and real-time interaction improvements. This demonstrates OpenAI’s progress in bringing more powerful AI models to market.

Alibaba’s Qwen research team has released the Qwen2.5-Coder Series, a collection of open-source (Apache 2.0 licensed) large language models (LLMs). The standout model, Qwen2.5-Coder-32B-Instruct, is claimed to match the coding capabilities of GPT-4o, despite being significantly smaller. This allows it to run efficiently on devices like the MacBook Pro M2 with 64GB of RAM. This development signifies a shift towards more accessible and powerful AI coding tools, potentially democratizing advanced coding capabilities for a wider range of users and developers.

The OWASP (Open Web Application Security Project) has released new security guidance for organizations running generative AI tools. The updated OWASP Top 10 for LLM focuses on addressing the growing threat of deepfakes, providing recommendations for risk assessment, threat actor identification, incident response, awareness training, and various event types. Additionally, the guidance advocates for establishing centers of excellence for gen AI security to develop security policies, foster collaboration, build trust, advance ethical practices, and optimize AI performance. This new guidance highlights the increasing need for a more comprehensive approach to securing AI and machine-learning tools, as attackers leverage AI to create more sophisticated and advanced threats.

The STORM AI research system, developed by Stanford University, leverages Large Language Models (LLMs) to conduct research using perspective-guided question asking in simulated conversations. Initially designed for web-based research, STORM now supports searching local document vector stores, allowing users to analyze internal data sets like the United States Federal Emergency Management Agency (FEMA) disaster preparedness and assistance documentation. This powerful research system mimics human research methods by iteratively refining document retrieval and synthesis, ultimately generating rich articles that can be used by humans in their pre-writing research.