MiniMax ships 428B MoE model M3.1 Flash at $0.15 per million input tokens

MiniMax confirmed on October 19 that the anonymous model "Space Bunny," which topped daily token-call-volume rankings on OpenRouter and OpenCode within three days of its late-September launch, is an early version of M3.1 Flash. The company's official account posted a video captioned "Welcome back, Space Bunny" to reveal the connection. MiniMax launched M3.1 Flash, a 428-billion-parameter mixture-of-experts model priced at $0.15 per million input tokens, with open weights due October 23. The company said the model delivers "a generational leap" in agentic tasks, coding, computer use, and cybersecurity. MiniMax said M3.1 Flash nearly triples online inference speed and cuts usage costs by 70%. Space Bunny processed more than 200 trillion tokens while operating anonymously. During that period, it joined DeepSeek V4.1 Flash and GLM 5.3 Flash in the top three on the same volume rankings. A Pro version of M3.1 Flash is also set to launch soon, according to MiniMax. DeepSeek released an update to its own V4.1-Flash model on October 17, reducing task execution time. MiniMax Group Inc's stock has dropped roughly 80% from its peak, even as DeepSeek pursues a raise of at least 800 billion yuan.

(Source：Vector Wire)