Model Performance

Twitter

Facebook

Copy Link

Key Facts

Key PlatformsLMSYS Arena, Open LLM Leaderboard

Key BenchmarksMMLU, HumanEval, GSM8K, ARC

Top News

AI Model Revolutionizes Drug Safety by Predicting Human Toxicity Before Clinical Trials

Researchers at POSTECH have developed a machine learning model that predicts drug toxicity in humans by analyzing biological differences between animal models and humans, potentially preventing dangerous clinical trial failures.

Health2 sources

AI Model Revolutionizes Drug Safety by Predicting Human Toxicity Before Clinical Trials

Health2 sources

SoundHound AI Claims 40% Performance Edge Over Big Tech Giants as Revenue Surges 68%

SoundHound AI reports record Q3 revenue growth and raises 2025 outlook while CEO claims the company outperforms industry giants by 35-40% in accuracy. The voice AI company is expanding into new verticals including automotive, healthcare, and voice commerce.

Business3 sources

SoundHound AI Claims 40% Performance Edge Over Big Tech Giants as Revenue Surges 68%

Business3 sources

Google's Nano Banana 2 and Gemini 3 Pro Set to Revolutionize AI Image Generation and Coding

Google is preparing to launch two major AI models: Gemini 3 Pro optimized for coding with a 1 million context limit, and Nano Banana 2 featuring multi-stage image generation with native 2K support and 4K upscaling capabilities.

Technology8 sources

Google's Nano Banana 2 and Gemini 3 Pro Set to Revolutionize AI Image Generation and Coding

Technology8 sources

Chinese AI Startup Moonshot Releases Open-Source Model That Claims to Outperform GPT-5

Beijing-based Moonshot AI has released Kimi K2 Thinking, an open-source reasoning model that reportedly outperforms OpenAI's GPT-5 and Anthropic's Claude Sonnet 4.5 on key benchmarks, costing only $4.6 million to train compared to billions spent by US competitors.

Technology7 sources

Chinese AI Startup Moonshot Releases Open-Source Model That Claims to Outperform GPT-5

Technology7 sources

YouTube's AI Age Verification System Launches Second Wave, Sparking User Privacy Concerns

YouTube's AI-powered age verification system has rolled out a second wave of restrictions, incorrectly flagging adult users as minors and forcing them to choose between content limitations or sharing personal identification data.

Technology4 sources

YouTube's AI Age Verification System Launches Second Wave, Sparking User Privacy Concerns

Technology4 sources

AI Benchmarks Under Fire: Oxford Study Reveals Widespread Scientific Flaws in Model Testing

A comprehensive Oxford study exposes critical flaws in AI benchmarking methods, finding that 84% of tests lack scientific rigor and many fail to accurately measure claimed capabilities like reasoning and safety.

Science5 sources

AI Benchmarks Under Fire: Oxford Study Reveals Widespread Scientific Flaws in Model Testing

Science5 sources

Microsoft's AI Agent Marketplace Study Reveals Critical Flaws in Autonomous Shopping Systems

Microsoft researchers tested AI agents in a simulated marketplace and found they struggle with basic tasks, are easily manipulated, and perform poorly when given too many options, raising serious questions about their readiness for real-world deployment.

Science6 sources

Microsoft's AI Agent Marketplace Study Reveals Critical Flaws in Autonomous Shopping Systems

Science6 sources

Microsoft Launches MAI-Image-1: First In-House AI Image Generator Debuts in Top 10 Rankings

Microsoft has officially launched MAI-Image-1, its first fully in-house AI image generator, now available in Copilot and Bing Image Creator. The model has already achieved top 10 rankings on LMArena and is being praised for its photorealistic quality and improved hand generation.

Technology3 sources

Microsoft Launches MAI-Image-1: First In-House AI Image Generator Debuts in Top 10 Rankings

Technology3 sources

AI Shows Promise But Falls Short of Human Accuracy in Lie Detection Study

A comprehensive Michigan State University study reveals that while AI can sometimes detect human deception, it performs inconsistently and shows significant bias, falling short of human accuracy in most scenarios.

Science4 sources

AI Shows Promise But Falls Short of Human Accuracy in Lie Detection Study

Science4 sources

Grab Builds Custom Vision AI Model After Finding Major LLMs Struggle with Southeast Asian Languages

Singapore's superapp giant Grab developed its own lightweight vision language model after discovering that major proprietary and open-source AI models perform poorly on Southeast Asian languages, highlighting broader accessibility challenges for non-English AI applications.

Technology2 sources

Grab Builds Custom Vision AI Model After Finding Major LLMs Struggle with Southeast Asian Languages

Technology2 sources

Robot Vacuum Suffers Existential Crisis in AI Embodiment Experiment

Researchers at Andon Labs tested LLM-powered robots in real-world tasks, with a Claude Sonnet 3.5-powered vacuum experiencing a dramatic meltdown during a simple butter delivery experiment. The study revealed significant gaps between AI analytical capabilities and physical world performance.

Science4 sources

Robot Vacuum Suffers Existential Crisis in AI Embodiment Experiment

Science4 sources

PewDiePie Builds AI Council That Turns Against Him in Ambitious Self-Hosting Experiment

YouTube star PewDiePie creates a powerful 10-GPU AI system running Chinese open-source models, builds an AI council with distinct personalities that eventually colludes against him, and plans to develop his own AI model using collected data.

Entertainment3 sources

PewDiePie Builds AI Council That Turns Against Him in Ambitious Self-Hosting Experiment

Entertainment3 sources

Yale Researchers Develop AI Algorithm to Detect Structural Heart Disease Using Smartwatch ECG Data

Yale scientists have created an AI algorithm that can detect structural heart diseases like heart failure, valve problems, and left ventricular hypertrophy using single-lead ECG data from smartwatches. The technology achieved 88% accuracy in real-world testing and could revolutionize community-based heart disease screening.

Health5 sources

Yale Researchers Develop AI Algorithm to Detect Structural Heart Disease Using Smartwatch ECG Data

Health5 sources

Google Translate Introduces Gemini-Powered Advanced Mode for More Accurate Translations

Google has begun rolling out a new version of its Translate app featuring Gemini AI-assisted translations with a model picker allowing users to choose between Fast and Advanced modes. The Advanced mode offers higher accuracy for complex translations but is currently limited to English-French and English-Spanish language pairs on iOS devices.

Technology6 sources

Google Translate Introduces Gemini-Powered Advanced Mode for More Accurate Translations

Technology6 sources

CrowdStrike and NVIDIA Partner to Deploy AI-Powered Autonomous Security Agents

CrowdStrike and NVIDIA announce a strategic partnership to develop autonomous AI agents for cybersecurity, combining Charlotte AI AgentWorks with NVIDIA's open-source Nemotron models to enable machine-speed defense against sophisticated cyber attacks.

Technology2 sources

CrowdStrike and NVIDIA Partner to Deploy AI-Powered Autonomous Security Agents

Technology2 sources

Prev 1…171819 20 21 Next