A mid-2026 chart says trillion-parameter models have become the global baseline
dralandthompson · x · 2026-07-29
A newsletter post visualizes the mid-2026 “1T+ AI model club” and argues that trillion-parameter models have become the new global standard.
Key points from the post and chart:
- It says China has 11 current trillion-parameter models available versus 8 in the US.
- The chart groups major models by lab, including Fable, GPT-6, Opus, GPT-5, Gemini 3.x Pro, Grok 4.x, K3, M3 Pro, ERNIE 5.x, Qwen 3.8 Max, MiMo V2.x Pro, and others.
- The author notes that each model can cost hundreds of millions of dollars to train.
- The memo frames this as evidence that trillion-parameter systems are now the established baseline worldwide.
More from Infra
- Meta's commitments jumped to $349B in a quarter as server memory prices tripled ahead of Oct 21 earnings — tengyanAI · 2026-10-16
- Meta releases Muse Spark Open, trained with vLLM-powered RL inference from day one — colinmcnamara · 2026-10-16
- Meta's open Muse Spark needs 493GB of memory — far beyond any consumer laptop — TejasKumar_ · 2026-10-16
- CoreWeave plunges 20% as credit markets price >60% default odds; is this dot-com all over again? — kimmonismus · 2026-10-16
- MiniCPM-V 4.7 35B-A3B hits 17+ tokens/s on iPad with ~3GB memory — techNmak · 2026-10-16
- China rolls out 'token factories' and a 2026-28 token economy action plan — nordicinst · 2026-10-16