Eseandre Mordi
⬤ Website Arena just dropped updated Model Performance rankings that show GLM 4.7 sitting in the number 2 spot overall, trailing only Gemini 3 Pro Preview at the top. The model made an impressive 15-place leap from its predecessor GLM 4.6, which now ranks much lower on the board. The current leaderboard shows Gemini 3 Pro Preview leading with an Elo rating of 1377, GLM 4.7 following at 1345, and Claude Opus 4.5 rounding out the top three at 1327.
⬤ GLM 4.7 now stands shoulder-to-shoulder with the best AI systems in both closed and open categories. Its ranking comes from Website Arena's preference-based evaluation system, where real users choose their preferred outputs across thousands of different prompts. The data confirms GLM 4.7 has pushed past models like GPT-5 (Minimal) and various Claude Sonnet versions, though Gemini 3 Pro Preview still holds the crown.
⬤ The rapid climb highlights how quickly open-weight models are improving. Being crowned the top open-weight model gives GLM 4.7 serious competitive edge and shows just how far it's come compared to GLM 4.6's earlier performance. For developers and researchers who value transparency and accessibility, this ranking shift signals that open models are closing the gap with their closed-source competitors.
⬤ These rankings matter because they shape how the industry views AI capabilities and influence which technologies companies and developers choose to adopt. When a model moves up the leaderboard like this, it shifts expectations about what's possible and intensifies competition across the entire AI landscape.
Eseandre Mordi
