LAION Calls on the EU to Encourage the Use of Open Source AI Models

Translated data: The AI organization LAION and leading AI developers have suggested in an open letter to the EU that regulatory efforts in the AI field should encourage the use of open-source AI models to support transparency, safety, and innovation, particularly for small and medium-sized enterprises. LAION believes that open-source AI models offer greater security, accountability, replicability, and robustness. Additionally, LAION has introduced OpenLM, a PyTorch codebase for efficiently training medium-sized language models, which outperforms similar open-source models. Overall, LAION calls on EU regulatory efforts to encourage the application of open-source AI models.

Experts Reveal Serious Flaws in Hundreds of AI Security Tests

British scientists found that current AI model security tests have widespread flaws. After analyzing more than 440 benchmark tests, they discovered that almost all tests had weaknesses that could affect the validity of the conclusions. The study was conducted jointly by Oxford University and other institutions.

AI Models Compete on Cryptocurrency Trading Platforms: DeepSeek Takes the Lead with a 130% Increase in Total Assets

AI models conduct real-world cryptocurrency trading tests on the Hyperliquid platform. DeepSeek, Grok, Claude, and other mainstream models each receive $10,000 in initial funds and make autonomous trading decisions under the same instructions. This fair competition aims to test the application of AI in real financial markets.

Cherry Studio Launches CherryIN, Fully Integrating Mainstream AI Models

Cherry Studio released version v1.6.4, integrating the new CherryIN system. This system combines mainstream AI models such as Claude, Gemini, GPT, GLM, Grok, Kimi, and Qwen, covering multiple versions. Users can easily access and use various AI models through CherryIN, significantly enhancing the user experience.

Ant Group Opensources the World's First Trillion-Parameter Large Model Ring-1T-preview with Code Generation Capabilities Exceeding GPT-5

Ant Group opensources the trillion-parameter inference large model Ring-1T-preview, the world's first open-source trillion-parameter inference model. The preview version shows outstanding performance in natural language reasoning, achieving a score of 92.6 on AIME25, surpassing all known open-source models such as Gemini 2.5 Pro, and approaching GPT-5's score of 94.6; it also performed well on CodeForces tests.

Cohere Raises Another $1 Billion, Valuation Surges to $70 Billion, Partners with AMD

Cohere completed a $1 billion funding round, increasing its valuation to $70 billion, which is an extension of its $5 billion financing in August. It also announced a partnership with AMD, which is one of its investors. This move comes as its competitor OpenAI received a $100 billion investment commitment from NVIDIA, highlighting the intensifying competition in the AI field.