GLM-5.2 tops BridgeBench reasoning leaderboard
Zhipu AI's newly released GLM-5.2 model secured the top position on the BridgeBench reasoning leaderboard, significantly outperforming competitor Kimi K2.7 Code. The benchmark data also indicates a reasoning regression in Kimi K2.7 compared to its predecessor Kimi K2.6, highlighting the challenge of maintaining steady reasoning improvements.
While GLM-5.2 showcases impressive reasoning performance on paper, the real story here is the regression in newer model iterations like Kimi K2.7, proving that version upgrades do not always guarantee improvements in complex reasoning.
DISCOVERED
2h ago
2026-06-15
PUBLISHED
2h ago
2026-06-15
RELEVANCE
AUTHOR
bridgemindai