Artificial intelligence systems may be good at generating text, recognizing images, and even solving basic math problems—but when it comes to advanced mathematical reasoning, they are hitting a wall.

Recent advances in Vision Language Models (VLMs) have shown significant progress in mathematical reasoning, yet they still face a critical bottleneck with problems that require visual assistance, such ...

Large-scale language models with long CoT reasoning, such as DeepSeek-R1, have shown good results on Olympiad-level mathematics. However, models trained through Supervised Fine-Tuning or Reinforcement ...

Epoch AI, the developer of a mathematics benchmark, did not initially disclose funding from OpenAI due to a non-disclosure agreement, and this only became known when OpenAI set a new record on the ...

AI coding assistants are benchmarked mostly on Python and JavaScript. If you maintain a framework outside that bubble, you've probably seen AI tools generate code that looks plausible but uses ...

Large language models (LLMs) have gained significant attention in solving planning problems, but current methodologies must be revised. Direct plan generation using LLMs has shown limited success, ...