Side-by-side comparison · models
GLM 4.5 vs GLM 4.5 Air (Thinking)
What we have checked about each, line by line. Every value shows where it came from and when we last confirmed it.
Models
GLM 4.5
GLM-4.5 is our latest flagship foundation model, purpose-built for agent-based applications.
Models
GLM 4.5 Air (Thinking)
Efficient GLM model for fast reasoning, coding, and agent workflows.
Checked facts
Line by line
| Measure | GLM 4.5 | GLM 4.5 Air (Thinking) |
|---|---|---|
| Input price | $0.600 per million tokens | Not checked yet |
| Output price | $2.20 per million tokens | Not checked yet |
| Price on z-ai | $0.110 per million cached input tokens | Not checked yet |
| Price on z-ai | $0.600 per million input tokens | Not checked yet |
| Price on z-ai | $2.20 per million output tokens | Not checked yet |
| Latest version | z-ai/glm-4.5 | Not checked yet |
| Time to first token (ms) on z-ai | 4,552 | Not checked yet |
| Output speed (tokens a second) on z-ai | 25 | Not checked yet |
| Time to first token (ms) | Not checked yet | 2,390 |
| Output speed (tokens a second) | Not checked yet | 82.8 |
| Tokens served on OpenRouter in a day | 217.7M (for 10 Nov 2026) | Not checked yet |
| Tool calls on OpenRouter in a day | 1.2K (for 10 Nov 2026) | Not checked yet |
| Artificial Analysis Intelligence Index | 12.8 | Not checked yet |
| LMArena text rating (Elo) | 1,430 | Not checked yet |
| Design Arena rating (Elo) | 1,166 | Not checked yet |
| Lech mazur writing | 73.4% (for 3 Aug 2025) | Not checked yet |
| Weirdml | 40.6% (for 5 Aug 2025) | Not checked yet |
A line appears only when at least one record has a value we confirmed from a source. Values we have not confirmed are left out, never guessed. Records of different types share fewer lines.
Add a third
Compare with one more record
How this works
Share it, or build your own
This page works for any two or three records in the directory, of any type. The link stays the same whichever order you name them in, so you can share it as it is. Change the records in the address, for example /directories/vs/claude-code-vs-cursor, or start from our editorial comparisons.