MOE: what I’ve logged
MOE comes up in 2 posts of my AI log, from October 2026 to November 2026, and each item below links to the primary source. The sources come from Thinking Machines and inclusionAI, filed under open weights, frontier and science.
Items that mention MOE
- Thinking Machines 20 Nov 2026
Welcome Inkling 2 Mini
A multimodal MoE with about 26B total and 3B active parameters, a 1M token context and a looped middle of the decoder, with day-zero support in Transformers, llama.cpp, MLX and SGLang.
- inclusionAI 14 Oct 2026
Ling-3.1-flash
A 560B hybrid-linear MoE with 25B active parameters and a 1M-token window, MIT licensed, with day-0 SGLang support and multi-token-prediction decoding.