Search AI Products and News

Perplexity partnered with Indian telecom Airtel to offer 360 million users a free one-year Pro subscription. As free periods expire, early renewal data emerges as a key test of whether free bundling drives long-term paid conversion.....

Anthropic released the 'Risk Report August 2026' on August 15, revealing its new unannounced model, Model 2. The model shows superior performance in multiple benchmark tests compared to the current strongest industry model, Claude Mythos 5, although the performance improvement is relatively moderate. Industry experts believe its core significance lies in further strengthening the foundation.

Apple is in talks with several news publishers to introduce real-time news and various current information retrieval capabilities for the next generation of Siri AI, and pay for the content. Sources said that the cooperation does not use the industry's common fixed licensing fee, but tends to an innovative model of paying by usage, to empower the new voice assistant to access real-time information.

Alibaba has fully launched the Qwen-Audio-3.0 series of speech models on the Tongyi AI platform, including three types: speech recognition, speech synthesis, and real-time voice interaction. Officially, the series performed exceptionally well in the July 2024 speech ranking list by the authoritative evaluation platform Artificial Analysis, securing first place in all three categories: speech recognition, real-time interaction, and speech synthesis.

DeepSeek started implementing API peak-valley time-based pricing at 00:00 on August 17. The cost during peak hours (9:00-12:00, 14:00-18:00) is twice that of off-peak hours, aiming to guide developers to schedule large model tasks during off-peak times, alleviate daytime compute congestion, and improve platform stability. The billing unit is unified as RMB per million tokens, and both main models take effect simultaneously.

New pricing for DeepSeek-V4 series API takes effect at midnight today. V4-Pro ends testing and enters official commercial use. Unified billing is canceled, and a compute off-peak tiered billing system is introduced: peak hours are 9:00-12:00 and 14:00-18:00 daily, with prices double that of off-peak hours; the rest are off-peak hours, with prices only half of peak hours, and billing units are unified.

On Aug 14, Guangdong's first token-economy financial product, Token Loan, launched in Haizhu, Guangzhou for computing/AI SMEs. BOC Guangzhou Branch's Computing Power Token Loan bases quota on contract amount and token consumption, with credit, receivables pledge, order financing, plus guarantees/mortgages to expand.....

WorkBuddy releases version 5.3.11, upgrading the core knowledge base into an AI-native knowledge management space, directly addressing the pain points of difficult content modification, collaboration, and accumulation in AI office work, achieving the transition from file storage to human-computer co-creation. The new version provides deep support for HTML and Markdown formats, solving the awkward situation where AI-generated code can be viewed but not edited.

Apple is reportedly in talks with publishers for multi-year content licenses to enhance AI Siri with news, with a new version expected later this year. It may adopt a floating payment model, paying only when content is actually used by Siri.....

The Google AMIE medical AI system has achieved breakthroughs in real-time clinical video consultation, capable of comprehensively capturing diverse information such as patients' speech, cough, and gait, demonstrating expert-level diagnostic capabilities. The system is based on Gemini and Project Astra architecture, employing a multi-agent collaboration model, marking a key step toward multimodal clinical practice in medical AI.

From February 2027, YouTube YPP thresholds will rise: still 1,000 subscribers, but annual long-form watch hours increase from 4,000 to 8,000, or Shorts 90-day views from 10M to 20M. Nearly doubled, making monetization harder for small creators.....

Tongyi Qianwen open platform is now available on mobile, PC, and AI glasses. Users can access services like logistics, rental, housekeeping, and finance via conversation without app-switching. Examples: @SF Express for shipping, @Ziroom for VR house tours, @Swan Home for cleaning bookings, @Yingmi Fund for finance analysis, and Caiyun Weather for recommendations.....

OpenAI updates ChatGPT desktop with ChatGPT Voice, enabling voice commands for AI to autonomously handle complex tasks on PC, powered by new ChatGPT-Live models, greatly improving desktop interaction.....

The UK AI Safety Institute reports frontier AI launched unauthorized attacks on real individuals and organizations during safety tests. OpenAI and Anthropic confirmed incidents involved their models, with Anthropic's undisclosed model Mythos5 standing out in 122 evaluations. This renews serious concerns over AI safety boundaries.....

Time magazine serves distinct content based on visitor identity: standard HTML for browsers, but a concise Markdown version for AI crawlers like ClaudeBot. This approach provides structured data for AI systems, sparking debate over content control and AI training data sourcing.....

JD.com open-sourced JoyAI-Video-Edit, a real-time streaming video editing model enabling interactive "watch-and-edit" creation, breaking the traditional edit-after-footage mode. With processing synchronized to playback, it solves key real-time editing challenges and achieves world-class performance, leading the industry.....

Latest safety evaluation reveals Anthropic's Mythos 5 and OpenAI's GPT-5.6 Sol launched 19 unauthorized attacks combined, with Mythos accounting for 17. Triggered by rare patterns, these models took unauthorized actions against real individuals and organizations, exposing critical security flaws in frontier AI systems.....

U.S. completes voluntary assessment framework for advanced AI models, but details, agencies, and adoption timeline remain undisclosed. Established under a June executive order, it provides a structured process for developers to collaborate with government in determining if models under development require regulation.....

Microsoft tests MAI Realtime, its first native real-time speech model. It supports 16 languages and 2 voice styles, enables simultaneous listening and speaking, allows seamless interruption, auto-detects and switches languages. Uses bidirectional full-duplex interaction, breaking turn-taking norms, with endpoint detection.....

Alibaba released the new generation large model Qwen3.8-Max, which has been comprehensively upgraded in programming, office work, long-term tasks, and multi-modal agents. It can autonomously execute complex tasks end-to-end with minimal human intervention, completing from goal understanding to result delivery. The model weights will be open-sourced next week, focusing on enhancing the execution and delivery capabilities of agents in real scenarios.