{"type":"rich","version":"1.0","provider_name":"Transistor","provider_url":"https://transistor.fm","author_name":"AI News Today | Julian Goldie Podcast","title":"China’s NEW Meituan LongCat 2.0 Tested!","html":"<iframe width=\"100%\" height=\"180\" frameborder=\"no\" scrolling=\"no\" seamless src=\"https://share.transistor.fm/e/47aafee3\"></iframe>","width":"100%","height":180,"duration":459,"description":"LongCat 2.0 (Open Source) Tested: Benchmarks, Games, and GLM 5.2 Comparison\nThe episode covers the official release of LongCat 2.0, an open-source Chinese agentic model revealed as the model behind the AoAlpha free API, with features like Sparse Attention, Zero Compute Experts, and MIPD. The host reviews benchmark claims (including Terminal Bench 2.1 and SWE-Bench Pro comparisons versus GPT-5.5 and Opus 4.8) and shares hands-on tests building game demos such as Dragon Realm, a Skyrim-style open world, and VoxelCraft, noting mixed results and frequent bugs. Access issues are mentioned, including difficulty using the API without a Chinese setup, so the model is tested via the website chat. A key point is that LongCat was trained on China’s Meituan chips without NVIDIA. Overall, GLM 5.2 is judged stronger in side-by-side game benchmarks, and the host promotes the AI Profit Boardroom and Agent OS setup.\n00:00 LongCat 2.0 Launch\n00:36 Benchmarks and API Hurdles\n01:38 Game Demos Dragon Realm\n02:23 Goldy Bench Verdict\n02:43 Trained Without NVIDIA\n03:32 How to Use It\n03:51 Eval Results vs GPT\n04:17 GLM 5.2 Showdown\n06:13 Final Take and Recommendation\n06:35 Agent OS and Boardroom Plug\n07:37 Wrap Up","thumbnail_url":"https://img.transistorcdn.com/Tp0JSUlyiAe2ran44b13cjk4ImQYS1QEMKCAa9Q7go0/rs:fill:0:0:1/w:400/h:400/q:60/mb:500000/aHR0cHM6Ly9pbWct/dXBsb2FkLXByb2R1/Y3Rpb24udHJhbnNp/c3Rvci5mbS9lNTYz/NGQxMTU2ZTM1MzU5/MDNiYjcwZjJmYTY5/ODJmOS5qcGc.webp","thumbnail_width":300,"thumbnail_height":300}