Google Bard GemPro Surpasses GPT-4, Claims Second Place in LMSYS Chatbot Rankings


On December 22, Zhipu Huazhang released and open-sourced the new generation large model GLM-4.7. The model has shown outstanding performance in multiple international benchmark tests, especially in the coding field, with comprehensive performance surpassing GPT-5.2. It ranked first in both open-source and domestic models on the authoritative coding evaluation platform Code Arena, focusing mainly on programming scenarios.
OpenAI released GPT Image 1.5, significantly enhancing ChatGPT's visual creation capabilities. The model can generate high-fidelity images, accurately understand and follow user instructions, support creating from scratch or editing based on uploaded photos, while retaining the core elements of the original image, with particularly excellent detail processing.
DeepSeek-V3.2 releases both standard and inference-optimized versions, supporting seamless switching across web, app, and API. It introduces a novel integration of 'thinking process' with 'tool calling', enabling parallel modes for extended reasoning and efficient task execution.....
AI fine-tuned with just two books mimics authors' styles, outperforming human imitators in evaluations by 159 participants, including experts.....
OpenAI recently sparked controversy due to secret model switching. Paid users reported that their GPT-4/5 were automatically replaced with low-computing power filtered models gpt-5-chat-safety and gpt-5-a-t-mini without prior notice, especially causing a sharp decline in response quality when dealing with sensitive content. This move has been questioned by users for infringing on their right to choose and be informed, highlighting the issue of insufficient platform transparency.