All News · AI Models & Launches
630 audio stories · Page 14 of 32
Krea 2 LoRAs Launch: Boosts AI Image Generation
Krea 2 LoRAs are now widely available to everyone, boosting AI image generation. This means broader fine-tuning capabilities are accessible.…
MiniMax M3: Unverified Claims & China's Intelligence Law
MiniMax has launched its M3 foundation model, claiming it's the first open-weight system to combine frontier coding-agent performance, a one…
Tether AI Open-Sources TurboQuant: 5x LLM Memory Savings
Tether AI has open-sourced TurboQuant, a new optimization framework. This system dramatically reduces large language model memory usage by u…
Alibaba's New AI Chip & LLM: Stronger, Cheaper AI
Alibaba has announced significant upgrades to its artificial intelligence offerings. The company is rolling out a more powerful AI chip and …
xAI's Grok Imagine: Language Drives Video AI Progress
An AI Research Engineer, Ethan He, states that much progress in visual intelligence comes from advancements in language models. This trend i…
Nvidia GTC Taipei: New AI Models for Robots & Cars
Nvidia is making a significant push into physical AI, unveiling new models for robots, autonomous vehicles, and video systems at GTC Taipei.…
Tether AI's TurboQuant: Data Center Memory for Your Device
Tether's AI Research Group has released TurboQuant, an open-source implementation of a Google Research memory compression algorithm. This te…
AI's Web3 Reality Check: Leading Models Fall Short
New research indicates that leading AI models are falling short in critical Web3 applications. A study by DMind AI, with researchers from Zh…
MiniMax M3 AI: Outperforms GPT-5.5, Cuts Costs
MiniMax, a Chinese AI start-up, has launched its new M3 AI model, which can outperform OpenAI's GPT-5.5. The company introduced M3 as part o…
Decart Leads AI World Models: $3.1B Valuation, 1M Users
Three startups are leading the way in developing "world models," seen as the next frontier in artificial intelligence. These models are mult…
NVIDIA Alpamayo 2: Robotaxi AI Model for Safer Driving
NVIDIA has launched Alpamayo 2 Super, their most powerful open reasoning model to date. This 32-billion-parameter model is designed for safe…
Piedmont Global's Connexus: AI Interpretation & Language Access
Piedmont Global has launched two new platforms: CensIQ and Connexus. CensIQ is a population intelligence and language access planning platfo…
NVIDIA RTX Spark: 1 Petaflop AI PC Power Arrives
NVIDIA has unveiled its RTX Spark superchip, significantly advancing AI PC capabilities. This new chip boasts 1 petaflop of computing power,…
MiniMax M3 AI: Faster, Cheaper Coding for Complex Tasks
MiniMax, a Chinese artificial intelligence start-up, has unveiled its new flagship AI model, M3. This model is designed for long and complex…
Nvidia Cosmos 3: Physical AI Model Accelerates Autonomous Systems
Nvidia has launched Cosmos 3, an open-world foundation model for physical AI. This new model uses a "Mixture of Transformers" architecture. …
NVIDIA Cosmos 3: New AI Sees, Hears, Plans Actions
NVIDIA has launched Cosmos 3, an open frontier foundation model for physical AI. This new model uses a mixture-of-transformers architecture.…
Agnes AI: Singapore-Built Models Rank Top 10 Globally
Agnes AI, a company based in Singapore, has achieved a top 10 AI lab placement on the Artificial Analysis global benchmark leaderboard. This…
DMind AI: No AI Safe for Web3's Highest-Stakes Tasks
No AI system is currently ready for Web3's highest-stakes tasks. This is the verdict from the first peer-reviewed Web3 AI benchmark, which t…
LeCun's World Model Proved: Current AI Brittle
Yann LeCun's world model concept, backed by $1.03 billion, now has a clearer theoretical and empirical mapping. New research defines when hi…
Sarvam AI: Vision Platform Price Cut After Rapid Adoption
Sarvam AI has cut the price of its document intelligence platform, Sarvam Vision. This move follows rapid adoption by developers and enterpr…