New security benchmark reveals major AI models including GPT and Gemini fail most jailbreak tests with scores as low as 40%, while Anthropic's Claude dominates safety rankings with 75-80% success rates, exposing widespread vulnerabilities across the industry despite advances in model size and capability.
OpenAI launches ChatGPT Image 1.5 with improved text clarity and photo realism, but the new model still lags behind Google's Nano Banana Pro in speed, accuracy, and overall image quality despite representing significant progress from previous versions.
OpenAI launches ambitious 'Superalignment' program targeting 2027 to solve AI safety crisis by creating human-level automated researchers that can align superintelligent systems with human values, though controversies remain over whose values should guide AI and who should make these critical decisions.
Startup Neurometric launches an AI leaderboard benchmarking 12 models across eight business tasks, revealing that smaller, task-specific AI models frequently outperform larger rivals in speed, cost, and performance, challenging the industry assumption that one large model can handle all workloads.
OpenAI CEO Sam Altman announces plans for a massive $1.4 trillion data center investment while promising major AI breakthroughs beyond GPT-5.2 in Q1 2025, as the company focuses on enterprise growth and competing with Google.
Tiiny AI launches the Guinness World Record-verified smallest personal AI supercomputer that runs 120-billion-parameter language models completely offline in a pocket-sized device weighing just 300 grams, delivering GPT-4o level intelligence with total privacy and consuming only 65W of power.
Google's Gemini AI agents mimic bee swarm behavior to boost clustering algorithm performance by 60%, with scout, employed, and onlooker artificial bees autonomously exploring data patterns through natural language prompts, achieving 0.820 accuracy on benchmark tests despite facing API connection and prompt compliance challenges.
Anthropic launches Agent Skills, an open standard that trains AI assistants through reusable workplace modules, as Fortune 500 companies deploy it for legal and finance tasks while tech giants Microsoft, OpenAI, and Stripe integrate the platform that transforms specialized AI agents into one universal assistant.
Sam Altman's AGI masterclass goes viral as industry experts debate potential AI bubble concerns, while new VoiceAgentBench research exposes significant gaps in current voice assistants' ability to handle complex tasks, even as breakthrough Voila voice-language models promise emotionally expressive real-time AI agents.
Google launches Gemini 3 Flash AI model, delivering Pro-grade reasoning capabilities 3x faster than previous versions while using 30% fewer tokens, now processing over 1 trillion tokens daily and becoming the default model in Gemini app globally with developer access through multiple platforms.