Switch Transformer by hand ✍️ ~ 13 steps walkthrough below
The Switch Transformer, by Fedus, Zoph, and Shazeer in 2022, is one of the papers that made sparse Mixture of Experts practical at scale. Today, frontier models use MoE to pack enormous parameter counts while activating
Astra burned $571 to beat portal
Deepseek just dropped a model that costs a fraction of a cent per million tokens Astra: tokens burned: enough to hit $571 Plus window: gone in -15 minutes quota feel: 4x hungrier than Sol DeepSeek V4.1 Flash: job: coding + agents near Opus 5
Proud to be a Day 0 host for DeepSeek V4.1 Flash.
Congrats deepseek_ai 🚀 🎁 50% extra credit on recharge, limited time 1M context · vision · tool calling $0.40/M in · $1.60/M out · $0.08/M cached Highest on 3 of 4 in your chart. Try on RunBios today:
Deepseek V4.1 flash is extremely good at designing landing and hero.
0.03 dollars was the total cost for this
You have reached the end of the archive
All of deepseek