← November 6, 2026 briefing · added in the 20:00 KST update
Ant Group releases Ling-3.2-flash, a native multimodal MoE model with 560B total and 25B active parameters, 1M-token context and two weeks of free API access (report)
Verified (fact)
Ant Group has released Ling-3.2-flash, a model designed to be multimodal from the ground up, according to Pandaily. It has 560B total parameters, with 25B active during inference. It supports a context length of up to 1 million tokens. API pricing is $0.30 per million input tokens and $0.90 per million output tokens, and access is free for the first two weeks after launch. The report does not say whether the weights will be released.
Why it matters (interpretation)
A large MoE multimodal model from a Chinese big tech company is now priced at $0.30 per million input tokens. This could intensify price competition among long-context multimodal APIs.
Confidence medium · official source pending
Sources
More about Ant Group
-
"Traces of export-controlled NVIDIA Blackwell found in six Chinese open-weight models," based on weight initialization (community report)
According to an analysis shared on GeekNews, the weights of six models released by DeepSeek, Moonshot AI, MiniMax and Ant Group show...
-
Ant Group's robotics model LingBot reaches 2.2 million downloads in nine months; drug-picking robots deployed at Guoda pharmacies in Shanghai (report)
According to Pandaily, LingBot, the robotics model Ant Group released as open source, has surpassed 2.2 million downloads and 33,000...