Orion Star Releases Open Source OrionStar-Yi-34B-Chat Model for Free Commercial Use


GLM-5.1 advances AI capabilities, handling 8-hour complex projects independently. It excels in coding and long-range tasks, outperforming peers on benchmarks like SWE-Bench Pro by fixing high-difficulty bugs.....
Apple team surpassed leading large models in UI design by improving open-source models. Traditional AI-generated code performs poorly in UI design because reinforcement learning with human feedback is too crude. Apple achieved a breakthrough with fine-tuning, enabling a small model to excel in specific tasks and solving the long-standing issue of interface development for developers.
European AI company Mistral AI released the new open-source coding model family Devstral2, including a 123B parameter flagship version and a 24B lightweight version, along with a complementary command-line tool Mistral Vibe CLI that supports automated programming. The model achieved 72.2 points on the SWE-bench benchmark, approaching the performance of top closed-source models, and the API is currently freely available, offering strong support for developers.
Step-Audio2mini, an open-source audio model by Step Fun, achieves SOTA in benchmarks. It excels in speech understanding, translation, and emotion analysis with unified audio inference and generation.....
Recently, Tencent's Hunyuan Video Model has officially begun recruiting testing partners on Platform X, marking a critical testing stage for this cutting-edge AI video generation technology. According to official sources, there is a high probability that the model will be open-sourced after the testing concludes, contributing its technological achievements to the global AI community. The Hunyuan Video Model is an important innovation by Tencent in the field of AI video generation, boasting over 13B parameters, making it one of the largest video generation models among open-source models.