GLM 5.2 ranks third on ProgramBench, but the thread warns averages can mislead
jyangballin · x · 2026-07-23
GLM 5.2 has landed 3rd place on the ProgramBench leaderboard as the first open-weight model evaluated in the latest update. The discussion around the result also highlights a broader evaluation caveat: reporting only average pass rates can hide cases where a model is still failing a meaningful fraction of instances.
The thread argues that looking at the number of almost-resolved and fully resolved tasks is more informative than a single aggregate score, because a few stubborn failures can reveal major shortcomings even when the average looks strong.
Related event: GLM-5.2 Ranks Third on ProgramBench Leaderboard(3 posts)→
More from Models
- Daily AI brief: GPT-Live-1 in API, OpenAI pauses $200 Pro signups amid Astra demand — koltregaskes · 2026-09-11
- Same Echo Maze prompt, three frontier models: all passed visually but shipped the same hidden bug — eyishazyer · 2026-09-11
- Benchmark scores drop from 89% to 19% on new evals — how benchmaxxing breaks leaderboard trust — airesearch12 · 2026-09-11
- ChatGPT tells user their question is too hard and to 'accept dumber answers' — phido3000 · 2026-09-11
- Developer Building a Unified Leaderboard of All Model Benchmark Scores — airesearch12 · 2026-09-11
- Rumor claims Kimi faked performance by serving Claude; DeepSeek new model surprises in evals — realsohamparekh · 2026-09-11