GLM-5.3-Flash
#3 v Small & Edge modelyZhipu Ai · v5.3 · flash · 17× · naposledy 29. 8. 2026
82
Momentum
GLM-5.3-Flash is a Mixture-of-Experts model released by Z.ai (Zhipu AI) on August 26, 2026, featuring 320 billion total parameters and 18 billion active parameters. It is the first natively multimodal model in the GLM-5 series, featuring a hybrid sparse-plus-linear attention architecture and a context window of approximately 1 million Token. Released under MIT license with open weights on Hugging Face, it outperforms its predecessor GLM-5.2 at roughly one-tenth the cost and approaches Claude Opus 4.8 in Coding and Agent Benchmarks.
Vývoj momenta
01.06.30.08.
Vlastnosti
| Key Benchmark (%) | 63.4% on DeepSWE v1.1 (vs. 46.2% for GLM-5.2); 57 points on Artificial Analysis Intelligence Index |
| Context Window (Tokens) | 1,048,576 tokens (max output 131,072 tokens) |
| License | MIT License (open weights) |
| Multimodality | Natively multimodal: text, image and video input, text output |
| Platform | API (Z.ai API Platform), Hugging Face (weights), ZCode, GLM Coding Plan, Chat (chat.z.ai), self-hosting via SGLang, vLLM, TokenSpeed, KTransformers |
| Price per 1M Tokens | $0.15 input / $0.50 output (list price); launch promo through September 9, 2026: $0.075 / $0.25 |
| Release Date | August 26, 2026 |