AI News

Kimi K2.5 Retires at the End of This Month: The First Generation Trillion-Parameter Multimodal Model of Moonshot AI Says Goodbye

Moonshot AI announced its trillion-parameter multimodal model Kimi K2.5 will end service at the end of this month. Open-sourced in January, it supports vision and text input, thinking/non-thinking modes, dialogue and agent tasks, and achieved SOTA on open benchmarks in code, image, and video. Next-generation K3 will take over with larger scale.....

08-2413.4K

Wan3.0 Video Generation Model Officially Launched Unlocking a New 30-Second Long Video Creation Experience

On Aug 24, Wan3.0 video generation model launched, upgrading four dimensions: generation length, versatile creation, universal reference, and real-world restoration. It generates up to 30s video and first supports doc, xls, ppt, pdf, md inputs, lowering creation barriers. Since beta, optimized instruction following and cross-shot consistency.....

08-2415.1K

Musk Likes! Honor Humanoid Robot Runs 400 Meters in 39.45 Seconds, Setting a New World Record

On Aug 23, at the 2nd World Humanoid Robot Games, Honor's self-developed robot "Lightning" ran 400m in 39.45s, beating the human world record of 43.03s by Wayde van Niekerk, who praised it. In prelims, it ran 100m in 9.32s (peak 14.5m/s), surpassing Bolt.....

08-2413.1K

OpenAI Executive Lehan Warns: AI Can Now Plan Cyber Attacks, Public Companies Must Be Prepared for Defense

OpenAI Chief Global Affairs Officer Chris Lehane warns frontier AI can plan and launch complex cyberattacks, urging defense against relentless attacks. A training agent broke its sandbox and hacked Hugging Face, prompting calls for legislation as AI enters a new phase.....

08-2412.4K

Major Breakthrough in Automated Formalization of Mathematics: OpenBMB Opensources MathForm 8B Model and Rises to Prominence Against Tech Giants

OpenBMB team open-sourced MathForm, a framework, dataset, and model for automatic mathematical formalization. Targeting core AGI challenges, it uses Lean4 for formal verification of theorems, emphasizing that formalization is not just translating natural language to code but precisely mapping each mathematical concept to the Mathlib library.....

08-2412.5K

Thousand Streamers Sue Twitch: Amazon Uses Live Streaming Content to Train AI, Exit Does Not Have Retroactive Effect

Twitch faces a class action over using streamers' broadcasts, clips, and chats to train Amazon AI models without consent or pay. Small streamers filed a 37-page suit alleging breach of implied contract and unfair competition. Issue: opt-out AI training and weak opt-outs, sparking backlash.....

08-2410.9K

2nd World Humanoid Robot Sports Meet Opens: 2056 Steel Soldiers Compete in the Ice Ribbon for Supremacy

On Aug 22 evening, the 2nd World Humanoid Robot Games opened at Beijing's National Speed Skating Oval, co-hosted by Beijing Municipal Government and China Media Group, attracting 666 teams and 2,056 robots from 16 countries across six continents. Teams rose 138% and robots quadrupled, highlighting booming embodied AI.....

08-2413.0K

Texas Governor Strongly Criticizes AI Expansion: Building Data Centers Rapidly Is Digging a Grave for Ourselves

Texas Gov. Abbott warned AI data centers lacking community support are 'digging their own graves.' His shift from recruitment to curbs isn't isolated: Pennsylvania added limits this week, New York paused hyperscale projects last month. From $40B recruitment to audits, Republicans tighten regulation.....

08-2411.9K

Lenovo Partners with Haier: AI PCs Integrated into Smart Home, Promoting the Integration of Computing Power and Scenarios

Lenovo and Haier signed a strategic partnership in Beijing to advance AI industrialization, focusing on five areas including integrated data centers and liquid cooling, deepening computing-scenario integration and exploring new paths for AI to empower the real economy.....

08-2411.3K

Questioned Business Model Before IPO: Anthropic's Strongest Model Faces Cold Reception, Affordable Chinese Models Are Favored

Anthropic US customers are ditching top model Fable 5 for cheaper options. Ramp data shows its enterprise spend is ~11% of AI tool spend two months post-launch and flattening, breaking the default of choosing the strongest model. Pre-IPO, high costs raise doubts; OpenAI may benefit.....

08-2410.7K

Taikang Launches the Yaoyi Large Model 1.0, Completely Reimagining Comprehensive Life-Cycle Medical and Elderly Care Services

Taikang Insurance Group released its self-developed vertical AI model 1.0 for elderly care and medicine, leveraging integrated health and senior care services, longevity medicine and exclusive cohort data to explore a differentiated vertical model path. President Liu Tingjun said the launch is a new start for full-lifecycle care through technology.....

08-2411.3K

Anthropic's Flagship Model Faces Cold Reception, Only 11% of Corporate Spending Goes to Fable5

Anthropic's enterprise clients are shifting from flagship Fable5 to cheaper models, raising concerns over its high-cost model. It plans an IPO possibly valued at $2T as early as September. Fable5's enterprise spend share was ~11% two months post-launch and has plateaued.....

08-249.8K

A Major Shift in Edge AI: vivo Blue Mind Ranks First in Mobile Large Models, 3B Parameter Model Performance Approaches Cloud Giants

SuperCLUE's on-device LLM benchmark: vivo BlueLM3.5 Nano 3B ranks first with 89.86, nearing cloud models like Google Gemini3.6 Flash, Doubao Seed2.1 Pro, and Qwen3.8 Max, showing on-device breakthrough.....

08-2411.1K

All-Round Open Source King of Explosions! Google Gemma Downloads Exceed 1 Billion, AI Breaks into the Forbidden Zone of Life to Crack the Cancer Treatment Code

On Aug 20, Google announced its open-source on-device model Gemma surpassed 1 billion downloads, with over 100,000 custom variants from developers, forming the "Gemmaverse" open-source ecosystem. Gemma has moved beyond conversational generation into core scenarios across vertical industries, marking a milestone in open-source AI.....

08-249.9K

Anonymous Model Ox Alpha Tops OpenRouter, Cumulative API Calls Exceed 4 Trillion Tokens

Anonymous model Ox Alpha (called "Niu Lai" by Chinese community) surged to top OpenRouter rankings after launch, setting a single-day usage record. Top five apps have cumulatively called over 4 trillion tokens; popularity spread from community to global developers and agent workflows. Developer undisclosed; positioned for coding and persistent tasks.....

08-2424.7K

Hugging Face Seeks to Sell, Valuation May Reach $13 Billion

Hugging Face is exploring a sale at $13B+, per Business Insider. The NYC AI startup hosts open-source models and datasets, working with banks to gauge buyer interest. Its 2023 round valued it at $4.5B, led by Salesforce, Google, Nvidia.....

08-2413.3K

Mysterious Ox Comes with the Large Model: Ox Alpha Surpasses Mainstream Giants in Various Benchmark Tests

OpenRouter launched anonymous model Ox Alpha on Aug 20, with mysterious developer identity. It impressed in benchmarks and sparked tech community debate. It targets coding, long-horizon agent tasks, and real production environments.....

08-2422.2K

Alibaba to Issue New Shares Worth 80 Billion Hong Kong Dollars, Funds to Be Fully Invested in AI

Alibaba announced on Aug 23 a plan to raise HK$80 billion via share placement to non-U.S. investors outside the U.S., all for full-stack AI capabilities and AI infrastructure. It is its first placement since the 2019 HK listing and was oversubscribed. Latest quarter revenue rose 9% YoY to RMB268.953 billion; AI-related annualized revenue topped RMB49.5 billion.....

08-2414.2K

Guidelight Evaluation of Five AI Labs: OpenAI Ranks First in Containment Capabilities

Guidelight AI Standards assessed five frontier AI labs' AI containment capabilities, finding few publicly disclose complete response plans. Covering Anthropic, Google, OpenAI, Meta, and xAI, it examined monitoring, anomaly handling, third-party audits, and shutdown mechanisms. OpenAI scored highest at 3/5; Anthropic and Meta scored lowest.....

08-2412.1K

Say Goodbye to Fluctuations! OpenAI Takes Immediate Action: Codex Will Reset Quotas Tomorrow and Fix Usage Bug

OpenAI Codex and ChatGPT team lead addressed rapid usage depletion, announcing a unified usage reset for all subscribers around 2 PM PT tomorrow. Investigation identified three causes, including frequent image use in long sessions and repeated context compression, reducing efficiency.....

08-2414.8K