AI News

512 articles, updated daily

Microsoft Ships Its Strongest Phi 4 Models Yet: Three Reasoning Models Rival Bigger Systems at a Fraction of the Size

Microsoft Ships Its Strongest Phi 4 Models Yet: Three Reasoning Models Rival Bigger Systems at a Fraction of the Size

Read more →
Sam Altman's World Launches Orb Mini: A Smartphone-Shaped Eye Scanner for Proving You're Human

Sam Altman's World Launches Orb Mini: A Smartphone-Shaped Eye Scanner for Proving You're Human

Read more →
Court Files Reveal Meta Expects $1.4 Trillion in Generative AI Revenue by 2035 — and a 2024 GenAI Budget Over $900 Million

Court Files Reveal Meta Expects $1.4 Trillion in Generative AI Revenue by 2035 — and a 2024 GenAI Budget Over $900 Million

Read more →
Amazon Launches Nova Premier: Its Most Capable AI Model Yet, With a 1M-Token Context and a Distillation Mission

Amazon Launches Nova Premier: Its Most Capable AI Model Yet, With a 1M-Token Context and a Distillation Mission

Read more →
Cohere, Stanford, MIT, and Ai2 Accuse LM Arena of Helping Top Labs Game the Leaderboard

Cohere, Stanford, MIT, and Ai2 Accuse LM Arena of Helping Top Labs Game the Leaderboard

Read more →
Meituan's First Open-Source LLM Arrives: LongCat-Flash Packs 560 Billion Parameters and Near-50% Faster Inference Than DeepSeek-V3

Meituan's First Open-Source LLM Arrives: LongCat-Flash Packs 560 Billion Parameters and Near-50% Faster Inference Than DeepSeek-V3

Read more →
The World's 10 Most Advanced AI Robots: From Fedor's Space Flight to Sophia's Fame

The World's 10 Most Advanced AI Robots: From Fedor's Space Flight to Sophia's Fame

Read more →
Baidu MuseSteamer 2.0 Goes All-Access: The First Audio-Video Unified Model to Generate Multi-Character Speech on Film-Grade Footage

Baidu MuseSteamer 2.0 Goes All-Access: The First Audio-Video Unified Model to Generate Multi-Character Speech on Film-Grade Footage

Read more →
DeepSeek-V3.1 Launches: The Hybrid-Reasoning Model That's DeepSeek's First Step Into the Agent Era

DeepSeek-V3.1 Launches: The Hybrid-Reasoning Model That's DeepSeek's First Step Into the Agent Era

Read more →
Apple's FastVLM Gives AI Eyes: The Hybrid Vision Encoder That Beats Token Pruning at Its Own Game

Apple's FastVLM Gives AI Eyes: The Hybrid Vision Encoder That Beats Token Pruning at Its Own Game

Read more →
Kuaishou Open-Sources Keye-VL 1.5: A New Benchmark for Video Understanding With 0.1-Second Temporal Grounding, 128K Context, and Cross-Modal Reasoning

Kuaishou Open-Sources Keye-VL 1.5: A New Benchmark for Video Understanding With 0.1-Second Temporal Grounding, 128K Context, and Cross-Modal Reasoning

Read more →
China's August AI App Rankings Reveal a Split Market: Beauty Apps Take the Cash While Chatbots Take the Users

China's August AI App Rankings Reveal a Split Market: Beauty Apps Take the Cash While Chatbots Take the Users

Read more →