Skip to content
Trending storyDeveloping

Meta's Muse Wonder Full 1.0 ranks fourth on Bug Hunt Bench

1 article1 sourcesince Oct 13Last article 1h ago ·

Overview

AISummary of 1 article

Alexandr Wang posts that Meta's Muse Wonder Full 1.0 scores 44 on Bug Hunt Bench at max effort, placing fourth among the models tested.

GPT-6 Astra scores 45 at max effort, ahead of Muse Wonder. Fable 5.1 scores 43 and Opus 5.5 scores 41.7.

The benchmark tests models on 105 bugs that frontier models missed in Q1 2026, drawn from two repositories.

Written by AI from the articles below · updated Oct 13, 1:23 PM ET

Check the sources:

Latest developmentOct 13, 1:19 PM ET
Muse Wonder Full 1.0 ranks fourth on Bug Hunt Bench, behind GPT-6 Astra

Article timeline

The articles in this story. Times are ET.

Oct 13
  1. Alexandr WangX
    Muse Wonder Full 1.0 ranks fourth on Bug Hunt Bench, behind GPT-6 Astra

    AIMuse Wonder Full 1.0 scores 44 on Bug Hunt Bench at max effort, placing fourth among the models tested. GPT-6 Astra at max effort scores 45, Fable 5.1 scores 43, and Opus 5.5 scores 41.7. The benchmark tests models on 105 bugs that frontier models missed in Q1 2026 across two repositories.

Heat trend

Not enough continuous observations to show a trend yet.