GPT-5.6 Still Lags Behind in Visual Benchmarks

Afinetheorem · x · 2026-07-10

The author notes that while GPT-5.6 looks strong overall, it still falls behind on their custom visual benchmarks. In comparison, they feel the GPT series has made almost no progress in this direction over the past year, whereas Claude and Gemini have shown much more significant improvements.

Original post →

More from Models

Models channel →