News
xAI introduced Grok 4 and SuperGrok Heavy subscription for $300
xAI launched Grok 4 with enhanced capabilities, expensive subscriptions, and a focus on safety and response accuracy
OpenAI strengthens protection against competitor espionage
OpenAI has implemented stricter security measures due to attempts by Chinese startups to copy its developments
Scandal Surrounding Grok: xAI’s AI Faces User Criticism
Grok by xAI sparked outrage due to biased and offensive responses after system changes, company promises to improve filters
ChatGPT tests the “Study together” feature for users
OpenAI is developing a mode where ChatGPT guides users through the learning process, asks questions, and checks answers.
Hidden AI Prompts Found in Scientific Preprints from Various Countries
Researchers from various universities masked AI prompts in their papers to receive favorable reviews.
Alibaba introduced the Qwen VLo model for step-by-step image generation
Qwen VLo by Alibaba analyzes, creates, and edits images based on text instructions, offering advanced capabilities for users
Grok 4 from xAI promises to become the new standard for developers
xAI skips Grok 3.5 and prepares Grok 4 with powerful capabilities for programming, text, and logical tasks
Huawei Open Sources Pangu Pro MoE Models
Huawei has provided open access to its Pangu language models, supporting text generation and information analysis
Tencent introduced Hunyuan-A13B with open source and flexible thinking
Tencent released Hunyuan-A13B — an open-source language model that allows changing the depth of reasoning during task execution
Website Owners Can Start Earning from AI Bot Access
Cloudflare launched a platform where publishers can control and monetize AI access to their content.
Meta chatbots will gain the ability to independently send messages to users
Meta has allowed its AI chatbots to initiate contact with users under certain conditions to increase engagement
Sakana AI Lab trains multiple language models to work together
Sakana AI introduced a technique that allows different language models to collectively solve tasks and reduce errors.
