MiniMax 在悉尼歌剧院搞了个会看观众反应自己改剧情的电影，12 月 7 日 NeurIPS 期间放映，挺少见玩法。
MiniMax 与创作者 @Pickford_AI 合作，在 NeurIPS 2026 期间于悉尼歌剧院放映生成式电影 Whispers。影片会根据现场观众的反应自动改写并重新生成画面。放映定于 12 月 7 日，为私人场次。
MiniMax × @Pickford_AI bring Whispers to the Sydney Opera House during #NeurIPS2026.
Cinema that rewrites and reanimates itself in response to its audience.
December 7 · Private screening
MiniMax 的 M3.1 Flash 用约 1/27 的价格在智能体自动化上跑赢了 Opus 5.5，还带 1M 上下文和视觉，做自动化流程的可以试试。
MiniMax 的 M3.1 Flash 在 SambaNova 云上跑出每秒 200+ tokens，速度比上一代快近 3 倍还便宜 70%，做 Agent 的可以试试。
MiniMax 的 428B 大模型被 Unsloth 压到 143.9GB 还能保持 91.5% 的输出一致率，想本地跑开源模型的朋友可以试试。
The non-obvious shift here is that a frontier Chinese model can now be customised inside a US platform, with fine-tuning runs staying in the customer's own Python loop rather than a black-box pipeline. Teams building agents can adapt M3.1 Flash with custom reward signals and environments via SFT or RL. This makes MiniMax a direct alternative to open-weight models fine-tuned on Together, with the model served alongside Western options.
Distillation and step reduction directly cut inference cost, which matters for teams serving video generation without large GPU fleets — the Apple Silicon and TPU ports show H3 is no longer tied to one hardware stack. Video developers and studios experimenting with open video models can pick the 4-step or 8-step variants for faster iteration. MiniMax (Hailuo) remains one of the few Chinese labs openly cultivating a third-party ecosystem around video generation.
MiniMax 把 M3.1 Flash 搬上了 Jim Keller 家的 Tenstorrent 芯片云，不靠 NVIDIA 也能跑开源权重模型，值得关注
