News from the AI & ML world
DeeperML - #azureai
|
Chris McKay@Maginative
//
OpenAI has launched a new series of GPT-4.1 models, including GPT-4.1, GPT-4.1 mini, and GPT-4.1 nano. These API-only models are not accessible via the ChatGPT interface but offer significant improvements in coding, instruction following, and context handling. All three models support a massive 1 million token context window, and they have a May 31, 2024 cutoff date.
GPT-4.1 demonstrates enhanced performance in coding benchmarks, surpassing GPT-4o by 21.4% on industry benchmarks. The models are also more cost-effective, with GPT-4.1 being 26% cheaper than GPT-4o and offering better latency. The GPT-4.1 nano model is OpenAI's cheapest model yet, priced at $0.10 per million input tokens and $0.40 per million output tokens. As a result of GPT-4.1's improved performance, OpenAI will be deprecating GPT-4.5 Preview on July 14, 2025. The GPT-4.1 series excels in several key areas, including coding capabilities and instruction following. The models have achieved impressive scores on benchmarks like SWE-bench Verified and Scale’s MultiChallenge, demonstrating real-world software engineering skills and enhanced adherence to requested formats. Several companies have reported significant improvements in their specialized applications, with GPT-4.1 scoring higher on internal coding benchmarks, providing better code review suggestions, and improving the extraction of granular financial data from complex documents.
Share:
References :
Classification:
Kailyn Sylvester@Microsoft Security Blog
//
Microsoft is making significant strides in the AI landscape, expanding its Copilot features to directly compete with leading AI models like ChatGPT and Gemini. These enhancements include web browsing capabilities, allowing users to task Copilot with booking tickets, making reservations, or shopping online. Furthermore, Copilot now boasts enhanced multimodal functionality, capable of analyzing live video feeds from mobile devices and responding to questions based on visual content. A key addition is the memory retention feature, enabling personalized interactions by remembering user preferences, although users retain control over managing or deleting these memories for privacy.
Microsoft is also investing in nurturing local AI start-ups through initiatives like the Cyberport x Microsoft AI Partnership Programme in Hong Kong. This program provides resources such as solution support, technical expert guidance, and business matching opportunities, aimed at enabling these start-ups to develop innovative AI solutions across various sectors. Six companies were selected to launch solutions on the Microsoft Azure Marketplace, addressing healthcare, insurance claims, risk management, and corporate sustainability. These companies showcased their technologies at a DEMO Day, attracting significant interest and fostering business collaborations. In addition to software advancements, Microsoft is pushing the boundaries of AI in gaming with WHAMM (World and Human Action MaskGIT Model). This generative AI model creates game visuals in real-time, representing a significant upgrade from its predecessor, WHAM-1.6B. WHAMM boasts faster visual output, generating images at over 10 frames per second with enhanced resolution. Trained on Quake II using intentionally curated data, WHAMM demonstrates enhanced performance in tracking existing environments and responding to user input, despite some limitations regarding stat accuracy, input lag, and context length.
Share:
References :
Classification:
Kailyn Sylvester@Microsoft Security Blog
//
Microsoft is actively enhancing its AI integration across Azure and Copilot, introducing new features and programs to support both enterprise security and AI innovation. The company is now offering the "Llama 4 herd" within Azure AI Foundry and Azure Databricks, providing users with more AI tools and resources. Simultaneously, Microsoft is working to improve Copilot's capabilities by integrating enhanced security measures and features designed to facilitate AI adoption within organizations. These advancements reflect Microsoft's commitment to making AI more accessible and secure for its users.
Microsoft is also working with external partners to foster AI development. Microsoft Hong Kong has collaborated with Cyberport to launch the "Cyberport x Microsoft AI Partnership Programme," aimed at nurturing local start-ups. This program provides benefits like solution support, expert guidance, and business matching opportunities to promising Hong Kong-based companies. Six companies were selected to participate, showcasing innovative AI solutions in healthcare, insurance, risk management, and corporate sustainability. In addition to external partnerships, Microsoft is focused on internal security enhancements related to AI. Microsoft Copilot utilizes classification labels as part of Microsoft Information Protection (MIP) to safeguard sensitive information, ensuring data security and regulatory compliance. These labels, applied manually, automatically, or suggested by Copilot, categorize data based on sensitivity levels, such as public, internal, or confidential. Furthermore, Microsoft is hosting a "Tech Accelerator: Azure Security and AI Adoption" event on April 22, 2025, designed to equip developers and cloud architects with the essential guidance and resources needed to securely plan, build, manage, and optimize their Azure deployments and AI projects.
Share:
References :
Classification:
|
BenchmarksBlogsResearch Tools |