GLM-5.3 Vision
#11 en Modèles multimodauxZhipu · v5.3 · vision · depuis 14. August 2026 (nur für GLM-5.3 text-only; "GLM-5.3 Vision" nicht erschienen) · 4× · vu le 24 août 2026
Important note: "GLM-5.3 Vision" does not exist as an officially released product from Zhipu/Z.ai based on extensive research. Multiple independent sources (official Z.ai documentation, tech media) consistently confirm that the actually released "GLM-5.3" model, launched August 14, 2026, is text-only and does not support native image input. A "GLM-5.3 Vision" was only requested by the community (poll by co-founder Jie Tang, June 2026) and remains officially unconfirmed and unreleased to date; anonymous leaked candidates ("Ox Alpha") are speculatively linked to it but not confirmed by Zhipu. Zhipu offers vision capabilities instead through separate models such as GLM-4.5V, GLM-4.6V, or the closed-source GLM-5V-Turbo. The values below therefore refer to the actually existing, officially docu
Fonctionnalités
| Key Benchmark (%) | CyberGym: 84.5% (GLM-5.3, no Vision score available) |
| Context Window (Tokens) | 1,000,000 tokens (max output 128,000 tokens) |
| License | Not officially confirmed; open weights for GLM-5.3 not yet published as of Aug 23, 2026 |
| Multimodality | No native image processing – GLM-5.3 is text-only (input & output); vision resides in separate models like GLM-4.6V/GLM-5V-Turbo |
| Platform | Z.ai Chat, Z.ai API (GLM Coding Plan, ZCode), OpenRouter |
| Price per 1M Tokens | $1.40 input / $4.40 output (GLM-5.3 API list price) |
| Release Date | August 14, 2026 (GLM-5.3, text-only; no Vision release) |