The iteration speed in the field of large-scale artificial intelligence models is continuously breaking industry records. On September 28th, the leading domestic AI company MiniMax officially announced that its new text model M3.1-Flash-Preview has been launched and is now open for full public testing.

Regarding core technologies and functional features, this model has two highlights: first, it fully supports native multimodal capabilities, and second, it has an ultra-long context window of up to a million characters. In practical application scenarios, it not only efficiently completes code bug fixes and complex function development from scratch, but also deeply participates in the entire closed-loop process of problem identification, code implementation, and test verification, providing developers with highly reliable productivity support.

Notably, before this model was officially unveiled, some keen developers had already linked it to a previous anonymous model called "Space Bunny" (meaning "Space Rabbit" in Chinese) that was launched on September 23rd. According to publicly available data, this mysterious "Space Rabbit" performed impressively during the Mid-Autumn Festival holiday, once ranking first in daily usage volume on the model aggregation platform OpenRouter and the programming platform OpenCode, attracting widespread attention from technology enthusiasts. With the official unveiling of M3.1-Flash-Preview, its underlying technical strength has also been comprehensively validated.