OpenAI has revealed its AI chatbot pricing strategy, charging up to $20,000 monthly for PhD-level research agents. Lower tiers cost $2,000 or $10,000, while ChatGPT Pro remains at $200. This approach targets specialized clients, showcasing confidence in AI's value despite challenges in finding high-paying customers. As AI evolves, demand for premium services may grow.

Google introduces AI Mode in Search Labs, using Gemini 2.0 for advanced AI responses to complex queries. Building on AI Overview, it targets detailed, multi-part questions via query expansion and broader searches. Initially available to Google One AI Premium subscribers, it will evolve with user feedback, adding visual and diverse formats. AI Overview also expands with Gemini 2.0 for coding and math in the US, removing login requirements and supporting teens. Google aims to address biases and en

Jimo AI launches the "Motion Simulation" feature, enabling users to generate dynamic videos where uploaded portraits mimic reference video actions with accurate emotional replication. Powered by ByteDance's hybrid-driven technology, it supports multiple aspect ratios and uses face motion tokenizer for realistic expressions. Users can choose from three templates or upload custom videos (max 30 seconds). Available on mobile and web, the platform ensures safety with content reviews and AI watermark

A US federal court denied Elon Musk's request for a preliminary injunction to block OpenAI's transition to a for-profit entity. Judge Yvonne Gonzalez Rogers ruled that Musk failed to meet the high standards required for such an injunction but agreed to expedite the trial. OpenAI defended its move as necessary for securing funds to advance AI technology, while Musk's lawsuit accuses OpenAI of betraying its non-profit mission. Microsoft, a major OpenAI investor, has not commented on the ruling. Th

Google added an AI-powered fraud detection feature in Android 15 to enhance the security of the "Messages" app by detecting scam messages in real time. It analyzes SMS, MMS, and RCS messages locally on devices to protect user privacy. Currently available in English for US, UK, and Canada, this feature aims to combat conversational text scams. Google also introduced real-time location sharing in "Find My Device," anti-theft tools, and shopping features in Chrome, along with games for Android Auto

OpenAI launches the NextGenAI Consortium with 15 institutions to drive AI advancements in education and research, backed by a $50 million investment. The initiative focuses on addressing high-impact challenges like healthcare and education, empowering students and researchers through grants, APIs, and computational resources. Founding members include Caltech, Harvard, MIT, and Howard University, among others. Initial projects explore AI applications in digital health, rare disease diagnosis, and

Cohere's research lab unveiled Aya Vision, an open-source multimodal AI model excelling in tasks like image captioning, text translation, and summarization across 23 languages. Available for free via WhatsApp and Hugging Face, Aya Vision aims to bridge performance gaps in multilingual multimodal tasks. It includes two versions, Aya Vision 32B and 8B, with competitive performance despite fewer resources. Trained using diverse datasets and synthetic annotations, it introduces AyaVisionBench, a new

Amazon is developing an advanced AI model under the Nova brand with hybrid reasoning capabilities, aiming for launch by June. This model will balance quick responses with complex problem-solving, competing with Google, OpenAI, and Anthropic. Amazon prioritizes cost-effectiveness, targeting a 75% price advantage over rivals, while aiming for top-5 performance in benchmarks like SWE and AIME. Led by Chief Scientist Rohit Prasad, this project intensifies competition with Anthropic, despite their $8

Google's Deep Research tool is rumored to develop two new features: "Show Thinking" for visualizing AI reasoning and "Create Audio Overview" for generating audio summaries. These enhancements aim to boost user experience and functionality, though they remain in development with no official release date. The updates promise smarter and more engaging interactions once launched.

Microsoft launched a native Copilot app for Windows, first available to Windows Insider members. Built with XAML, it features a sidebar and optimized UI, supporting image uploads, text generation, voice interaction, and conversation history. This update improves upon the previous web-based version and will roll out to all Windows 11 users soon.

Zhipu launched and open-sourced CogView4, an advanced image generation model under the Apache 2.0 license. It excels in complex semantic alignment, supports bilingual inputs, and generates high-quality images at various resolutions. CogView4-6B leads in benchmarks, reduces token redundancy by 50%, and integrates a multi-stage training strategy for aesthetic alignment. The model uses 2D RoPE and adaptive LayerNorm layers, replacing T5 with GLM-4 for bilingual support, enhancing applications in ad

Stanford researchers launched a free AI tool named Storm that generates high-quality research reports with 99% accuracy. Users input a topic on its website, and the tool creates detailed reports with citations, offering PDF downloads and real-time previews. It also has a GitHub repository for more details. This enhances research efficiency significantly.

Tongyi Lingma upgraded its model selection feature with DeepSeek V3, R1, and the new Qwen2.5-Max models. Qwen2.5-Max, pre-trained on 20 trillion tokens, excels in benchmarks like Arena-Hard and LiveBench, outperforming models such as DeepSeek V3 and GPT-4o. It ranks 7th globally in Chatbot Arena's blind tests and leads in math and programming skills. The platform now integrates Qwen2.5-Max for enhanced coding capabilities.

Samsung's new Galaxy A series, including the Galaxy A56 5G, A36 5G, and A26 5G, features AI-powered tools like Awesome Intelligence to enhance creativity and simplify tasks. Key highlights include Circle to Search, Best Faces, Object Eraser, IP67 durability, up to six years of OS updates, and Samsung Knox security. The devices offer vibrant color options, a 6.7-inch Super AMOLED display, and Digital Key functionality in select markets. Prices vary by region, ensuring accessible cutting-edge tech

Conch Video launched I2V-01-Director, an image-to-video model complementing the existing T2V-01-Director text-to-video model. Together, they complete the 01-Director series, marking the end of the Video-01 line and paving the way for Video-02. Both models are globally accessible via Conch Video and MiniMax platforms, offering advanced lens control with 15 camera movements to create over 100 effects. The new mobile app enables on-the-go video creation from text or images, featuring natural langua

Apple's AI strategy, Apple Intelligence, has faced slow rollout and underwhelming performance compared to rivals like OpenAI and Microsoft. Despite initial hype, issues such as low feature urgency, incorrect AI-generated news alerts, and technical delays have hindered progress. Apple plans to upgrade Siri with a unified core in future software updates, but a competitive conversational version may not launch until 2027. Facing talent poaching and GPU challenges, Apple is restructuring its AI depa

Anthropic PBC raised $3.5 billion in a late-stage funding round, valuing it at $61.5 billion. Led by Lightspeed Venture Partners, the Series E included Cisco Investments and Salesforce Ventures. The company unveiled Claude 3.7 Sonnet, an advanced AI model with customizable processing, and Claude Code, a tool for automating coding tasks. Proceeds will enhance AI systems, expand computational resources, and fund research into model interpretability. Anthropic is also expanding globally, opening of

Google's Wildlife Insights project, part of Google Earth Outreach, uses SpeciesNet to classify wildlife images with over 2,000 labels. SpeciesNet, trained on 65 million images, is now open-source on GitHub, enabling broader biodiversity monitoring. Other tools like Microsoft's PyTorch Wildlife framework also support automated camera trap image analysis.

Podcastle launched its AI text-to-speech model, Asyncflow v1.0, offering over 450 AI voices and an API for developers. With cost-efficient development and rapid voice cloning advancements, Podcastle competes with ElevenLabs and others in marketing, education, and content creation. Priced at $40 for 500 minutes, it aims to integrate audio, video, and AI tools on a unified platform. Founder Artur Yeritsyan highlights growing video processing demand alongside audio services.

T-Mobile announced a collaboration with Perplexity and other partners to launch an AI-powered smartphone under $1,000 in 2026. The device will feature Perplexity's assistant on the lock screen and T-Mobile's Magenta AI app for existing users. This move aims to leverage AI, redefining telecom's role beyond network services amidst stiff competition from Apple and Samsung. Perplexity is also expanding its capabilities to include proactive tasks like booking flights and sending emails. While success