34M-param Nerve-1 Runs on CPUs to Route AI Tasks, Claiming 30% Inference Cost Cut
saheedniyi_02 · x · 2026-10-15
trydecideAI has trained Nerve-1, its first from-scratch model: a 34M-parameter router trained on 23B+ tokens that decides how much reasoning effort a task needs and directs it to the right model. It runs entirely on regular CPUs with millisecond decisions, and the team estimates at least 30% inference cost savings. Rolling out soon to all users and businesses on the platform.
More from Infra
- OpenAI infra lead leaves to start up after locking in double-digit GWs of compute — DesireeCachette · 2026-10-15
- Who Sells Data directory maps 159 AI training data vendors, compares Scale vs Mercor — cneuralnetwork · 2026-10-15
- Compute is the next commodity: building financial infrastructure for GPU markets — AnneliesGamble · 2026-10-15
- OCP slide reveals NVIDIA Rubin Ultra NVL576 will use CPO for scale-up networking — BenBajarin · 2026-10-15
- TensorFold ships macOS app for local LLM inference, reports 874 tok/s on RTX PRO 6000 — HankYeomans · 2026-10-15
- Enroute open-sources Git server hitting 7,600 concurrent pushes/s per repo on Raft and S3 — DanielLockyer · 2026-10-15