BridgeBench rates Meta's Muse Wonder at 611, 149 points behind its leader
Matthew Miller's coding-model leaderboard places Muse Wonder eighth of 29 models; its score is an editorial assessment, not a success rate.
By Ryan Merket · Published · Updated
Primary source: X
Why it matters
BridgeBench's rating gives Meta's new coding model a comparative position, but the proprietary methodology and low security score make the axis-level results more useful than the overall number alone.
BridgeBench, the AI coding-model benchmark built by BridgeMind founder Matthew Miller (@matthewmillerai), rates Meta's new Muse Wonder at 611, putting it eighth among 29 models and 149 points behind the leaderboard's Claude Opus 5.5 leader. The rating gives Meta's first frontier-model launch in this cycle a comparison against coding rivals. BridgeBench says its scores are comparative assessments, not percentages of tasks completed successfully.…