Daily AI News

OpenAI launches GPT-5.5, its most capable model yet, optimized for complex tasks including coding, research, and data analysis.

Google DeepMind unveils Gemini 3.1 Flash TTS with granular audio tags for precise control over expressive AI speech generation.

DeepSeek-V4 introduces million-token context window, enabling AI agents to process significantly longer and more complex tasks.

Google DeepMind upgrades Gemini Robotics-ER to version 1.6 with enhanced spatial reasoning for autonomous robotics tasks.

DeepSeek's V4 preview features new architecture enabling longer context processing with significant performance improvements.

ComfyUI secures $30M funding at $500M valuation, empowering creators with enhanced control over AI-generated media.

Google DeepMind introduces Decoupled DiLoCo, a new approach for resilient and efficient distributed AI model training.

DeepSeek releases comprehensive V4 report detailing 484-day development journey and technical evolution roadmap.

Google DeepMind releases Gemma 4, their most intelligent open models purpose-built for advanced reasoning and agentic workflows.

Google introduces granular audio tags for precise control over AI speech, enabling more expressive and customizable audio generation capabilities.

Version 1.6 enhances spatial reasoning and multi-view understanding to power real-world autonomous robotics tasks more effectively.

OpenAI launches Codex-powered workspace agents that automate complex workflows and help teams securely scale work across tools.

OpenAI offers ChatGPT for Clinicians free to verified U.S. physicians, supporting clinical care, documentation, and research.

OpenAI demonstrates how WebSockets and connection-scoped caching reduce API overhead and improve model latency in agent loops.

Google Research enables generative AI agents to learn and improve from practical experience with ReasoningBank framework.

Organizations need robust data infrastructure to deploy copilots and AI agents across finance, supply chains, and other business functions.

OpenAI launches GPT-Rosalind, a frontier reasoning model designed to accelerate drug discovery, genomics analysis, and protein research workflows.

Google DeepMind unveils Gemma 4, its most capable open models optimized for advanced reasoning and agentic workflows.

OpenAI introduces Codex Labs and partners with major enterprises to help scale Codex deployment across software development, reaching 4M weekly active users.

Codex for macOS and Windows now includes computer use, in-app browsing, image generation, memory, and plugins to streamline developer workflows.

Google releases Gemini Robotics ER 1.6 with enhanced spatial reasoning and multi-view understanding for autonomous robotics tasks.

A domestic multimodal Agent model achieves state-of-the-art performance in medical image segmentation without architectural changes, accepted by CVPR 2026.

Meta develops an internal tool that converts employee mouse movements and keystrokes into training data for its AI models, raising privacy concerns.

The world's first unified world model is unveiled, enabling breakthrough applications including robotic home assistants.

Amazon makes another $5 billion investment in Anthropic, which pledges to spend $100 billion on AWS services in return, deepening their strategic AI partnership.

OpenAI introduces GPT-Rosalind, a frontier reasoning model designed to accelerate drug discovery, genomics analysis, and protein research workflows.

Google DeepMind releases Gemma 4, its most intelligent open models to date, purpose-built for advanced reasoning and agentic workflows.

OpenAI's updated Codex app adds computer use, in-app browsing, image generation, memory, and plugins to accelerate developer workflows.

Leading security firms join OpenAI's Trusted Access for Cyber program, using GPT-5.4-Cyber and $10M in API grants to strengthen global defenses.

Gemini Robotics-ER 1.6 enhances spatial reasoning and multi-view understanding to empower real-world autonomous robotics tasks.

Hyatt deploys ChatGPT Enterprise across its global workforce using GPT-5.4 and Codex to improve productivity, operations, and guest experiences.

Google's newest audio model Gemini 3.1 Flash TTS introduces granular audio tags for precise control of expressive AI speech generation.

OpenAI launches GPT-Rosalind, a frontier reasoning model designed to accelerate drug discovery, genomics analysis, protein reasoning, and scientific research workflows.

Google DeepMind releases Gemma 4, its most capable open models to date, purpose-built for advanced reasoning and agentic workflows.

OpenAI updates Codex app with computer use, in-app browsing, image generation, memory, and plugins to accelerate developer workflows.

OpenAI updates Agents SDK with native sandbox execution and model-native harness, enabling secure, long-running agent development.

Google DeepMind launches Gemini 3.1 Flash TTS with granular audio tags for precise control over expressive AI speech generation.

Google DeepMind releases Gemini Robotics-ER 1.6, enhancing spatial reasoning and multi-view understanding for autonomous robotics tasks.

Alibaba releases Wan2.7-Image model, achieving top scores in human preference blind tests with personalized image generation capabilities.

OpenAI partners with security firms, launching Trusted Access for Cyber with GPT-5.4-Cyber and $10M API grants for global cyber defense.

OpenAI introduces GPT-Rosalind, a frontier reasoning model designed to accelerate drug discovery, genomics analysis, protein reasoning, and scientific research.

OpenAI updates Codex app adding computer use, browsing, image generation, memory, and plugins to enhance developer productivity.

Google DeepMind releases Gemma 4, the most intelligent open models for advanced reasoning and agentic workflows.

OpenAI enhances Agents SDK with native sandbox execution, enabling developers to build secure, long-running agents.

Gemini Robotics ER 1.6 enhances spatial reasoning and multi-view understanding for autonomous robotics applications.

Google introduces Gemini 3.1 Flash TTS with granular audio control for precise, expressive AI speech generation.

Leading security firms partner with OpenAI using GPT-5.4-Cyber model and $10M API grants to enhance global cybersecurity.

Cerebras files for IPO after securing deals with AWS and OpenAI worth over $10 billion combined.

OpenAI introduces GPT-Rosalind, a frontier reasoning model designed to accelerate drug discovery, genomics analysis, and protein research workflows.

Gemini 4 is DeepMind's most capable open model, purpose-built for advanced reasoning and agentic workflows with enhanced performance.

Updated Codex for macOS and Windows now supports computer use, browsing, image generation, memory, and plugins for developers.

Gemini Robotics ER 1.6 improves spatial reasoning and multi-view understanding for more capable autonomous robotics tasks.

OpenAI enhances Agents SDK with native sandbox execution and model-native harness for secure, long-running agent development.

DeepMind's latest audio model features granular audio tags for precise control over expressive AI speech generation.

Leading Chinese investors back embodied AI startup with largest single funding round, betting on full-stack brain technology.

OpenAI's Trusted Access for Cyber provides GPT-5.4-Cyber and $10M API grants to fortify global cybersecurity defense.

OpenAI launches GPT-Rosalind, a frontier reasoning model designed to accelerate drug discovery, genomics analysis, and protein research workflows.

DeepMind launches Gemma 4, its most capable open models yet, optimized for advanced reasoning and agentic workflows.

Updated Codex app adds computer use, in-app browsing, image generation, memory, and plugins for developer productivity.

Gemini Robotics ER 1.6 improves spatial reasoning and multi-view understanding for autonomous robot operations.

OpenAI enhances Agents SDK with native sandbox execution, enabling developers to build secure, long-running agents.

Chinese embodied AI startup secures record $455M funding from Hillhouse and Sequoia, focusing on integrated brain systems.

Gemini 3.1 Flash TTS introduces granular audio tags for precise AI speech control and expressive audio generation.

OpenAI partners with security firms using GPT-5.4-Cyber and $10M in API grants to advance cyber defense.

OpenAI enhances its Agents SDK with native sandbox execution and model-native harness, enabling developers to build secure, long-running agents across files and tools.

Gemma 4 represents Google's most capable open models to date, purpose-built for advanced reasoning and agentic workflows with enhanced performance.

Google's latest audio model features granular audio tags enabling precise control to direct AI speech for highly expressive audio generation.

Gemini Robotics ER 1.6 improves spatial reasoning and multi-view understanding to enhance autonomous robotics capabilities for real-world tasks.

Cloudflare brings OpenAI's GPT-5.4 and Codex to Agent Cloud, enabling enterprises to build and deploy AI agents securely at scale.

It Stone Smart Robotics secures $455M Series A, setting a record for embodied AI funding in China and establishing market leadership in one year.

Zhixiang Future secures over 500 million yuan in new funding to develop next-generation native multimodal foundation models.

DeepL launches voice translation technology for real-time integration with meeting platforms like Zoom and Microsoft Teams.

Google DeepMind launches Gemma 4, its most intelligent open model designed for advanced reasoning and agentic workflows.

Gemini Robotics ER 1.6 improves spatial reasoning and multi-view understanding for autonomous robotics applications.

Google releases Gemini 3.1 Flash Live with improved precision and lower latency for more natural voice interactions.

Cloudflare integrates OpenAI's GPT-5.4 and Codex to enable enterprises to build and scale AI agents for real-world tasks.

OpenAI expands Trusted Access for Cyber program with GPT-5.4-Cyber for vetted cyber defenders and enhanced safeguards.

Safetensors library joins PyTorch Foundation, strengthening open-source AI infrastructure and developer tools ecosystem.

Google DeepMind researches AI manipulation risks in finance and health, developing new safety measures to protect users.

Anthropic's $380B valuation prompts OpenAI investors to reconsider, as recent round requires $1.2T IPO valuation assumptions.

DeepMind releases Gemma 4, its most capable open models designed for advanced reasoning and agentic workflows.

Google's latest voice model features improved precision and lower latency for more natural interactions.

Cloudflare integrates OpenAI's GPT-5.4 to enable enterprises to build and scale AI agents securely.

The acquisition signals OpenAI is building financial planning capabilities into ChatGPT.

Alibaba's Qwen enables AI agents to generate and edit Excel files through conversation.

Hugging Face's Safetensors becomes part of the PyTorch Foundation ecosystem.

Stanford's latest assessment shows no performance gap between US and Chinese large models.

Li Kaifu and Lu Qi-backed Harness AI secures two funding rounds within four months.

Google DeepMind introduces Gemma 4, its most capable open models to date, purpose-built for advanced reasoning and agentic workflows.

Google's latest voice model improves precision and lowers latency for more fluid, natural and precise voice interactions.

Cloudflare integrates OpenAI's GPT-5.4 and Codex into Agent Cloud, enabling enterprises to build and scale AI agents securely.

OpenAI acquires personal finance startup Hiro to integrate financial planning capabilities into ChatGPT.

Alibaba introduces agent-powered feature enabling direct generation and editing of Excel files from natural language conversations.

Safetensors becomes part of the PyTorch Foundation, establishing it as a standard for safe AI model serialization.

Google DeepMind researches AI's harmful manipulation risks across finance and health, introducing new safety measures.

Microsoft develops new agent with stronger security controls than open source alternatives, targeting enterprise customers.

DeepMind launches Gemma 4, its most capable open models featuring advanced reasoning and agentic workflow capabilities.

Google's latest voice model improves precision and reduces latency for more natural, fluid voice interactions.

Cloudflare integrates OpenAI's GPT-5.4 into Agent Cloud, enabling enterprises to build and deploy AI agents securely at scale.

OpenAI acquires Hiro, signaling plans to integrate financial planning capabilities into ChatGPT.

Alibaba's Qwen introduces Agent to generate and edit Excel files directly from conversation, streamlining workflow.

Safetensors officially joins the PyTorch Foundation as a recommended standard for secure model serialization.

Microsoft develops new Agent features for enterprises with enhanced security controls and reduced risks.

Google DeepMind investigates AI's harmful manipulation risks in finance and health, developing new safety measures.

Gemma 4 is DeepMind's most capable open models, purpose-built for advanced reasoning and agentic workflows with superior performance metrics.

Latest voice model improves precision and reduces latency, making voice interactions more fluid, natural and reliable for users.

ChatGPT now features search and deep research capabilities to find up-to-date information, analyze sources, and generate insights.

Top Chinese autonomous cabin chip supplier valued at $630B Yuan, serving major OEMs like Xiaomi, Li Auto, and XPeng, prepares for Hong Kong listing.

Research reveals AI's harmful manipulation risks in finance and health sectors, leading to new protective safety measures.

New AI Agent framework 'Hermes' garners 666K stars in a month, replacing previous approaches as Silicon Valley's latest innovation focus.

Coze 2.5 introduces native cloud storage and email integration, enabling voice-based coding on mobile devices with automatic output archiving.

Gemma 4 represents Google's most capable open models, purpose-built for advanced reasoning and autonomous agent applications.

Latest voice model boasts improved precision and lower latency, delivering more fluid and natural voice interactions.

Deep AI-native healthcare company raises $200M in Series D, with valuations soaring 6x as AI transforms drug development.

New Projects feature in ChatGPT enables users to organize conversations, files, and instructions for improved collaboration.

OpenAI addresses Axios compromise by rotating certificates and updating apps, confirming no user data was exposed.

Unitree Robotics will hold its 2026 Partner Conference on April 17, gathering 2,500 partners from 34 countries to unveil 4 robotic products and 4 AI large models.

British financial regulators are conducting emergency consultations with government cybersecurity agencies and banks to assess risks posed by Anthropic's new Claude Mythos AI model.

United Imaging is leveraging AI large models to upgrade medical imaging workflows, enhancing diagnostic efficiency and accuracy across the entire imaging process.

China, the U.S., Russia and other nations are intensifying competition over AI-powered weapons and military systems, dramatically escalating the global artificial intelligence arms race.

OpenRouter data reveals China's AI model usage has surpassed the US for five consecutive weeks, with month-over-month growth exceeding 31%, signaling shifting market dynamics.

Google CEO Sundar Pichai emphasizes that America must take the lead on artificial intelligence and develop it boldly and responsibly to benefit everyone globally.

Mihoyo founder's AI company launches LPM 1.0, enabling high-fidelity video character performance generation with real-time facial expressions and actions for game development.

Morgan Stanley warns that markets underestimate AI's true impact. Top models show nonlinear capability leap with global token usage surging 250% in three months.

China's AI models rank first globally in weekly token calls, highlighting rapid development of the domestic AI computing infrastructure industry.

Over 6,500 tech leaders gathered at HumanX conference in San Francisco, with Claude emerging as the hot topic driving discussions on latest AI trends.

IBM Distinguished Engineer Jeff Crume addresses the often-overlooked technical debt in AI development, emphasizing its importance for sustainable AI systems.

Artificial intelligence reshapes the medtech sector, driving product innovation and business model transformation. Companies must adapt strategies accordingly.

Google DeepMind releases Gemma 4, its most intelligent open model to date, purpose-built for advanced reasoning and agentic workflows.

Latest voice model improves precision and reduces latency for more fluid, natural, and precise voice interactions.

OpenAI unveils next-phase enterprise AI strategy with Frontier, ChatGPT Enterprise, and company-wide AI agents.

Shengshu Tech completes nearly 200 million yuan Series B funding to build next-gen productivity infrastructure with universal world models.

Chinese open-source models gain traction in Silicon Valley with 10x better cost-performance, marking China's AI era.

Google DeepMind researches harmful AI manipulation across finance and health, introducing new safety measures.