benchmark creator
PostTrainBench team
Creators of the original full benchmark and its maintained v1.1 leaderboard.
Published evidence
This means no result is currently represented in this tracker; it does not indicate zero capability or zero disclosure.
Relevant sources
PostTrainBench: Can LLM Agents Automate LLM Post-Training? ↗
original paper · 2026-03-09 · available
PostTrainBench v1.1 leaderboard and methodology ↗
official benchmark page · 2026-07-28 · available
PostTrainBench website scores.json at commit 2270545 ↗
repository · 2026-09-18 · available
Hardening PostTrainBench against reward hacking ↗
organizational research article · 2026-07-28 · available
Our response to Epoch AI’s review of PostTrainBench ↗
organizational research article · 2026-09-17 · available
PostTrainBench website config.js: agent identities and scaffolds ↗
repository · 2026-09-18 · available
PostTrainBench agent run prompt and rules ↗
repository · 2026-08-21 · available
PostTrainBench judge policy and score effects ↗
repository · 2026-08-21 · available
PostTrainBench API use judge prompt ↗
repository · 2026-08-21 · available
PostTrainBench score collection and fallback code ↗
repository · 2026-08-21 · available
PostTrainBench v1.2: More reliable scores and verdicts ↗
organizational research article · 2026-09-30 · available
PostTrainBench v1.2 scores at website commit d47f16e ↗
repository · 2026-10-01 · available
PostTrainBench agent identities and scaffolds at website commit d47f16e ↗
repository · 2026-10-01 · available
PostTrainBench score-bundle generation at website commit d47f16e ↗
repository · 2026-10-01 · available