On October 27, Liquid AI released the LFM3 family, including six models spanning language, vision, speech, and structured decision-making. Built on a 1.7-billion-parameter foundation with specialized versions around 2 billion parameters, the models can be customized on a single GPU and deployed on devices, private servers, or cloud. The text model API is priced at $0.08 per million input tokens and $0.20 per million output tokens.
MiniMax announced the open-sourcing of its MiniMax M3.1 Flash model weights, with 428B total parameters and 23B activated parameters. The model launched on October 19 and processed over 100 trillion tokens within 72 hours. MiniMax is partnering with chip vendors including Huawei Ascend, NVIDIA, and AMD, along with multiple cloud platforms, for day-one integration, with communities such as vLLM and Hugging Face supporting ecosystem adoption.
On October 20, 2026, Jefferies maintained a Buy rating and HK$533 target price on Minimax, citing the release of the M3.1 Flash model with significant capability gains in agentic tasks, programming, computer use, and cybersecurity, plus roughly 3x faster inference and 70% lower costs.
On October 19, MiniMax released the M3.1 Flash model, cutting costs on agentic tasks by 70% compared to the previous generation. API pricing is RMB 1 per million input tokens, RMB 4 per million output tokens, and RMB 0.1 per million cache reads. Model weights will be open on October 23.
MiniMax released its new Flash model MiniMax M3.1 Flash, achieving generational improvements in agentic tasks, coding, computer use, and cybersecurity, with nearly 3x faster online inference and 70% lower usage costs. Input is priced at 1 yuan per million tokens, output at 4 yuan, and cache read at 0.1 yuan. It is now available on MiniMax API, MPlan, and MiniMax Code, with weights to be released later.
On October 14, at MIPCOM in Cannes, Huace officially released the trailer for "Journey to the West: The Lost 500 Years," a fully AI-generated 3D animated series. Developed by Huace as part of its strategic partnership with Utopai Studios, the AIGC-driven premium series will have six 15-minute episodes, is expected to premiere in 2027, and may later explore a theatrical release.
MiniMax announced that its latest text model M3.1-Flash-Preview has been launched on the MiniMax Code platform, providing stable productivity for daily development, from bug fixing to full feature development. The company also issued a quota reset card to reset all users' Token Plan quotas, and introduced benefits such as double check-in points and free tokens.
<body><div>Xiyu Technology open-sourced its native multimodal flagship model MiniMax M3, with 428B total parameters and 23B activated parameters, the first of its kind in</div></body>
<body><div>AI startup MiniMax releases flagship model M3, achieving 59% in near-real software engineering tests, surpassing GPT-5.5 and approaching Opus 4.7, with million-</div></body>
MiniMax Releases M3 Flagship Model with 1M Context and Native Multimodal Capabilities