The supercomputing internet platform has recently upgraded its MaaS service, officially integrating the QwQ-32B reasoning large model. This model had already gained global attention for its outstanding performance in mathematical reasoning and code generation, being considered on par with the hundred-billion-parameter DeepSeek-R1 model.

After this integration, users can directly experience QwQ-32B’s multi-turn interaction and long-text parsing capabilities through the newly added Chat conversation entry on the platform. This feature aims to meet the needs of various scenarios such as academic research, code generation, and daily consultations.

In addition, the model API interface service of the supercomputing internet platform has seen new progress. Since its launch, the cumulative number of calls has quickly exceeded a million, demonstrating developers' strong interest in this service. To attract new users, the platform is now offering registration benefits: newly registered users can enjoy a free quota of one million tokens, allowing them to call models more efficiently and quickly validate ideas and requirements. This move aims to further explore the application potential of models in complex reasoning scenarios.