AI & RoboticsMore in AI & Robotics→
GLM-5.3-Flash pushes open AI on cost and chips
Key Takeaways
- Z.ai released GLM-5.3-Flash, a 320B-parameter multimodal model with 18B active parameters and a 1M-token context window.
- Artificial Analysis data in the source places the model three points behind GLM-5.3 while costing about 7.5 times less per task.
- Z.ai says the model ran on Chinese AI chips, highlighting a possible shift away from Nvidia dependence for advanced inference workloads.
DE
DT Editorial Team··via the-decoder.com