#LoopedMoE
Thinking Machines just dropped Inkling 2 Mini – a 26B looped MoE that nailed 27/28 benchmarks. Curious how a mini model outperforms the giants? Dive into the results! #Inkling2Mini #LoopedMoE #LLMBenchmarks

🔗 aidailypost.com/news/thinkin...
November 21, 2026 at 12:27 PM