NanoGPT Speedrun Frontier B
AI benchmark · Agentic
What is NanoGPT Speedrun Frontier?
Prime Intellect's frontier leaderboard for the nanoGPT optimizer speedrun. Autonomous coding agents are given the speedrun task and must improve the training recipe on their own; each model's best validated record is reported alongside the share of the gap between the 3,290 baseline and the 2,600 human record that it closed. Prime Intellect reports 153 autonomous runs across 18 frontier models, run under several agent harnesses (claude-code, codex, kimi-code, grok-cli, qwen-code, muse-code, prime-agent, pi), with an equal-budget comparison that replays each model's best run under a fixed 24-hour budget and open traces for most entries.
Tier B
Established: 15 or more models scored, or curated RL directories rate it.
How much stands behind a tier: usage, who backs it, a confirmed line from its own site or paper, and a clear read on X. High with three of the four, medium with two, low otherwise.
- 17 models scored
- in curated RL directories (1 pt)
How we got to tier B
- Adoption: 17 models scored. Tier B.
- Not yet read on X.
- In curated RL directories: 1 point, a floor of none.
- Not yet confirmed by a line from its own site or paper.
A clear view on X moves a tier one step up or down; most entries don't move. How tiers work.
Not on the map yet: no line from its own site or paper places it in a domain.
Next: browse all benchmarks. Entry from the RL Research daily scrape of public sources, 2026-10-11 snapshot.