Yuanxiang Open Source High-Performance Large Model XVERSE-65B-2 Base Version Enhancing Code and Math Capabilities

Meta象 has announced the open-source release of its high-performance large model, XVERSE-65B-2 Base Edition, with targeted enhancements in code and mathematical capabilities, while maintaining the strong generative creation, role-playing, and tool usage abilities of its predecessor. This model has surpassed GPT3.5 and Llama2-70B in comprehensive evaluations and has outperformed the open-source benchmarks comprehensively. In the latest SuperCLUE Chinese general-purpose large model comprehensive evaluation, XVERSE-65B has distinguished itself among domestic open-source models. Meta象 has open-sourced three large models, with XVERSE-65B being the earliest and largest parameter free commercial model open-sourced in China.

Xiaomi Large Model MiMo Public Testing Extended, Users Can Enjoy Free Experience Until 2026!

Xiaomi announced that the free public testing period of its self-developed large model MiMo-V2-Flash has been extended by 20 days, until January 20, 2026. The model has 309 billion parameters, with 15 billion activated parameters, and performs excellently in reasoning and code generation. This move aims to provide users with a longer experience period and demonstrate Xiaomi's continuous investment and confidence in the AI field.

OpenAI Releases GPT-5.2: The Next-Generation AI Assistant Empowering Daily Work

OpenAI released the GPT-5.2 series model, positioned for daily professional use, aiming to enhance user economic value. The series includes the Instant, Thinking, and Pro versions, with significant improvements over GPT-5.1 in tasks such as spreadsheets, presentations, code writing, long text understanding, and image processing.

NetEase Youdao Dictionary Reveals 2025 Annual Word - DeepSeek Leads with 8.67 Million Searches

The 2025 annual hot words of NetEase Youdao Dictionary have been revealed, with "DeepSeek" topping the list with 8.67 million searches, becoming the first annual word originating from a domestic AI large model. The search popularity rapidly increased after the release of the DeepSeek-R1 model in February, and subsequent technological breakthroughs also drove peak query periods. College students and working professionals are the main search groups, and users often extend their browsing to related concepts such as "large models" after looking up words, forming a chain of "looking up words - learning concepts," reflecting the trend of AI technology popularization driving public awareness deepening.

Alibaba Cloud Large Model Prices Cut in Half! Tongyi Qianwen 3-Max Call Cost Reduced by 50%, Only 10% Fee Charged for Cache Hits

Alibaba Cloud BaiLian announced that starting from November 13, 2025, the core call fee for the Tongyi Qianwen 3-Max model will be halved, and the cache billing strategy has been optimized, significantly reducing the cost of enterprise AI applications. This move aims to lower the entry barrier for large model usage and accelerate digital transformation for small and medium enterprises.

GPT-5.1 Officially Released! Not Only Smarter, but Also Empathetic: Added 6 Personality Styles, AI's First Emotional Dependency Safety Assessment

OpenAI released GPT-5.1, marking a new stage in the AI competition where emotional intelligence is integrated. The model focuses on enhancing emotional value, personalized interaction, and human-like expression, addressing user feedback about AI feeling cold. It uses a dual-model architecture: Instant mode for quick responses, and Thinking mode for deep thinking. It is being rolled out globally in batches, and paying users can continue using GPT-5 for transition within the next 3 months.