A cheap, natively multimodal GLM-5 model (320B total, 18B active) for coding and agent work. It has a 1M tokens context window, and its weights are openly available. Through Z.ai’s API it costs $0.15 per million input tokens and $0.50 per million output tokens.
Among the tools we track, it’s available in Cursor and Amp. Amp uses it for the low-cost end of its model range. It’s MIT-licensed, one of the most permissive licences among current open models.
What Z.ai reports
- Z.ai reports 63.4 on DeepSWE v1.1
- Z.ai reports 29.0 on Z.ai Code Bench v1.0 (run in Claude Code)
Vendor-reported figures, not independently verified by Code×Radar.
Related changes
- Amp replaces its Dial with a per-thread model selector · November 12, 2026
Sources
Facts last checked November 15, 2026.