Chat with the newest AI models to work on your hardest problems

Discuss your mathematics with state-of-the-art frontier models in parallel.

Free Sponsored Credits for researchers without other funding. Your chats are not used for AI training.

371
Users
1,950
Chats
5,878
Chat turns
11,220
Model answers
ModelAll models are queried through the business APIs of their companies. Answers keptHow often users continued the conversation with an answer of this model family (all versions of the model, such as all Claude Opus versions). Answer timeThe median time a model takes for an answer in the chat, at its default thinking effort. Median priceLog in to see the credit points a typical answer uses. Hosted inWhere the company that runs the model is based. Under its terms, data may also be processed in other countries. Retention periodHow long the company keeps a copy of your questions and the answers, mainly to detect misuse. Content flagged as a possible policy violation may be kept longer. The arrow next to each figure links to the company's own page. Model trainingUnder the companies' terms for business customers, chats sent through their business interfaces are not used to train models, and data sharing is switched off in our accounts.
Gemini 4 Neon 54% 26s USA 55 days No
GPT-6.2 Astra 96% 1m31s USA at least 30 days No
GPT-6.2 Sol 86% 1m00s USA at least 30 days No
Claude Opus 5.5 70% 1m07s USA up to 30 days No
GPT-6.2 Astra Pro premium model 95% 5m25s USA at least 30 days No
Claude Fable 5.2 premium model 74% 3m30s USA up to 30 days No
November 21, 2026 · The chat now runs the newest models from Anthropic, OpenAI and Google: Claude Fable 5.2, GPT-6.2 Astra, GPT-6.2 Astra Pro, GPT-6.2 Sol, and Gemini 4 Neon.
November 19, 2026 · For your students: MathTutor, our sibling project, is an AI tutor that discusses exercises and the students' own solution attempts, giving hints instead of finished solutions (Bastani et al., PNAS 2025). MathTutor is a pilot project at Ruhr University Bochum.
October 6, 2026 · Version 2 of our paper Benchmarks in Leipzig is now on arXiv. We believe that the current way of submitting benchmark problems has reached its limits, and we are looking into new ways of benchmarking. In the meantime, we are closing the submission system.