GLM-5.3-Flash
CodeFreemium

GLM-5.3-Flash

Z.ai’s MIT-licensed multimodal MoE — 320B total / 18B active, 1M-token context, at $0.15 per million input tokens.

4.5
Facts checked

About GLM-5.3-Flash

GLM-5.3-Flash (Z.ai, formerly Zhipu AI) is an MIT-licensed, natively multimodal Mixture-of-Experts model with 320B total parameters and 18B active per token, a 1M-token context window and image and video input. It previewed anonymously as the "Ox Alpha" stealth model before Z.ai claimed it on August 26, 2026. On the Artificial Analysis Intelligence Index v4.3 (captured September 8, 2026) it scores 42 at a measured $0.25 per index task — level with GPT-5.6 Terra (max, 42), ahead of Claude Sonnet 5 (max, 38) and GPT-5.6 Luna (max, 38), and behind Claude Opus 5 (max, 51) and GPT-5.6 Sol (max, 47) — at $0.15/$0.50 per million input/output tokens. Weights are on Hugging Face at roughly 306 GiB in FP8.

Tags

codingopen-weightmoemit-licensechina1m-contextmultimodal

Similar Tools