Skip to content
Trending storyMonitoring

Epoch AI gives Fable 5 and GPT-5.6 Sol 3000 GPU-hours each to beat GRPO post-training baseline

1 article1 sourcesince Oct 7Last article 3d ago ·

Overview

AISummary of one article

Epoch AI gave Fable 5 and GPT-5.6 Sol 3000 GPU-hours each to develop a new post-training technique intended to outperform the standard GRPO baseline.

The post reports the task setup but does not state the results.

Written by AI from one article, by Epoch AI

Check the sources:

Latest developmentOct 7, 2:08 PM ET
Epoch AI tests Fable 5 and GPT-5.6 Sol on novel post-training research

Article timeline

The articles in this story. Times are ET.

Oct 7
  1. Epoch AIOfficial
    Epoch AI tests Fable 5 and GPT-5.6 Sol on novel post-training research

    AIEpoch AI gave Fable 5 and GPT-5.6 Sol 3000 GPU-hours each to develop a new post-training technique intended to outperform the standard GRPO baseline. The post reports the task setup but does not state the results.

Heat trend

Not enough continuous observations to show a trend yet.