This model is Underrated, its in my opinion so much helpful as DS & GLM recent models

#16
by Oldbrain-exe - opened

Shows real capabilities on long running tasks (good memo, good focus, quite intelligent) and amazing result, good to mention that this was the "Space-Bunny"

220B params with 1-bit UD-IQ1_M 129 GB 23B activated per token, RIP Scam Ai from AMD Ryzen Strix Halo, Gorgon Halo, Nvidia DGX spark.

220B params with 1-bit UD-IQ1_M 129 GB 23B activated per token, RIP Scam Ai from AMD Ryzen Strix Halo, Gorgon Halo, Nvidia DGX spark.

What are you talking about?

MiniMax-M3.1-Flash is 220B params with 1-bit UD-IQ1_M 129 GB 23B activated per token

Sign up or log in to comment