Z.ai's GLM-5.3-Flash scores 57 on Artificial Analysis at $0.15 per 1M input tokens
Z.ai's open-weights GLM-5.3-Flash, released August 26, 2026, scores 57 on the Artificial Analysis Intelligence Index against a median of 27 for comparable open-weight models, at $0.15 per 1M input and $0.50 per 1M output tokens under an MIT license. The MoE design — 320B total parameters, 18B active, 1M-token context — puts frontier-adjacent reasoning and image input inside a self-hostable model that any team can deploy commercially without a vendor contract. Output runs 50.2 tokens per second, below average for the class, and the benchmark page is live so the figures will move.
Source: artificialanalysis.ai ↗
GLM-5.3-Flash scores 57 on the Artificial Analysis Intelligence Index, placing it well above average among other open weight models of similar size (median: 27).
Why this matters
- → Open-weights model scores frontier-adjacent reasoning at $0.15/1M input tokens—self-hostable without vendor lo
- → MoE design (320B total, 18B active) makes frontier capabilities deployable on constrained hardware
- → MIT license + 1M context window enable commercial use cases previously locked behind proprietary APIs