[GLM-5.3 Intelligence Index Soars to 60: Matches Kimi K3, API Pricing Unchanged] According to monitoring by Beating, Zhipu has officially opened the GLM-5.3 API. Previously, the API was not immediately launched when the model was released. Meanwhile, an independent evaluation by Artificial Analysis has been published: the Intelligence Index has surged from GLM-5.2's 53 to 60, matching Kimi K3 and only 1 point lower than GPT-5.6 Sol. Once the equal-weight parameters are publicly disclosed as planned, it will tie with Kimi K3 as the highest-rated open-weight model with an AA score. The most significant increase was seen in Agent. The Elo score for GDPval-AA v2 jumped from 1524 to 1770, an increase of 246 points, ranking second among all models, only behind Claude Opus 5's 1855, and over 100 points higher than Kimi K3's 1668. The API pricing has not increased. Input remains at $1.40 per million tokens, output at $4.40, and cached input at $0.26, exactly the same as GLM-5.2. However, GLM-5.3 is noticeably more token-efficient. AA testing shows that its average output per task is approximately 18,700 tokens, 20% more than GLM-5.2. As a result, the actual cost per task has risen from $0.44 to $0.68, an increase of about 55%. Even so, it is still cheaper than Kimi K3's $0.84 and GPT-5.6 Sol's $1.23 among models in the same tier. [Original Link]