List your company

Visual Reasoning on R-Bench-V Physics

71.3AccuracyMetacognitive Harness
10.35626.1784257.822May 13, 2026

Evaluation results

MethodLinks
Metacognitive HarnessBackbone=Sonnet-4.62026.0571.37
Sonnet-4.6Evaluation Mode=Single-pass (Pass@1)2026.0564.3—
GPT-5-miniLeaderboard Rank=Top-12026.0531.8—
o3Leaderboard Rank=Top-22026.0520.4—
o4-miniLeaderboard Rank=Top-32026.0512.7—