GLM-5.2 hits 8.5% on ProgramBench, ranks 3rd overall

parth007_96 · x · 2026-07-23

ProgramBench updated its leaderboard, and GLM-5.2 is now the first open-weight model they evaluated to reach 8.5% almost resolved, placing 3rd overall.

The chart also shows the top scores across the benchmark’s 200 tasks and 248,000+ tests, with GPT-5.5 (xhigh) still leading at 13.5% almost resolved and 0.5% resolved.

Original post →

More from Models

Models channel →