Atlas

Benchmarks

← All benchmarks

BridgeBench V3 Arena Overall

Indexes · 2026-07-10

The BridgeBench V3 Arena overall score averages confidence-weighted Elo across seven judged arenas. Each arena is weighted by matches divided by matches plus ten, and an unplayed arena contributes the 1000-point starting rating.

Top models (higher is better)

ModelScore
Claude Fable 51079
Opus 4.81000
GLM-5.21000
Qwen3.7-Max1000
Grok 4.5991
GPT-5.6 Sol934
Loading Atlas data…