Harel Lidar

Article summary · Sept 2026

How Frontier Labs Train LLMs

Mostly someone else’s words. I keep it as reading, not as my own writing.

Step back from the stages and the claim from the top of the post holds up: by 2026 there is **one recipe**, and the reports are variations on it. Here is the whole pipeline in one breath — _curate and dedup human-plus-synthetic data and choose the mixture by scaling-law forecasting; pre-train a RoPE/GQA/SwiGLU/RMSNorm MoE on a tokens-per-parameter ladder, deliberately over-trained, in FP8; mid-train on reasoning-dense data and extend context; SFT/cold-start to install behaviors and format; run GRPO-family RL with verifiable rewards, entropy control, and trainability filtering; align with a gated reward stack and an instruction hierarchy; measure with cheap NLL plus a contamination-guarded benchmark zoo; and red-team against a preparedness framework._ If you internalize that sentence, every report in the table below reads like a fill-in-the-blanks.

Architecture

Settled core

Data

Pre-Training

  1. Scaling Ladders - train a family of models varying across one single axis and fit the loss<>X curbe.
  2. [Maximal update parameterization tricks to transfer models across scales. ](https://arxiv.or g/abs/2203.03466)
  3. Small Scale instability proxies - reproduce the big-run faillure points (loss/attention spikes) by increasing the LR.

Post Training I - SFT & Distillation

SFT

Distillation

Post Training II - RL

  1. Verification as the ceiling on RL: How do we get high-quality verification that is both informative and scalable and does not suffer from hacking by a smart capable model that is situationally aware?
  2. The data dilemma: The high-quality human text is running out and we want to generate more data showcasing specific capabilities and patterns. The way to synthetically generate them is still an open question.
  3. The cost and fragility of long-horizon RL algorithms is an open question. How do we scale PRMs? How do we assign responsibility for success or failure