Side-by-side comparison · models
Deepseek R1 Distill Qwen 32B vs GLM-4.7-Flash
What we have checked about each, line by line. Every value shows where it came from and when we last confirmed it.
Classic open reasoning model for transparent math, coding, and deliberate problem solving.
Models
GLM-4.7-Flash
Efficient GLM model for fast reasoning, coding, and agent workflows.
Checked facts
Line by line
| Measure | Deepseek R1 Distill Qwen 32B | GLM-4.7-Flash |
|---|---|---|
| Input price | Not checked yet | $0.060 per million tokens |
| Output price | Not checked yet | $0.400 per million tokens |
| Price on cloudflare | Not checked yet | $0.060 per million input tokens |
| Price on cloudflare | Not checked yet | $0.400 per million output tokens |
| Price on novita | Not checked yet | $0.010 per million cached input tokens |
| Price on novita | Not checked yet | $0.070 per million input tokens |
| Price on novita | Not checked yet | $0.400 per million output tokens |
| Latest version | Not checked yet | z-ai/glm-4.7-flash-20260119 |
| Time to first token (ms) | Not checked yet | 1,020 |
| Time to first token (ms) on cloudflare | Not checked yet | 315 |
| Time to first token (ms) on novita | Not checked yet | 1,207 |
| Output speed (tokens a second) | Not checked yet | 122.5 |
| Output speed (tokens a second) on cloudflare | Not checked yet | 44 |
| Output speed (tokens a second) on novita | Not checked yet | 54 |
| Tokens served on OpenRouter in a day | Not checked yet | 1.3B (for 10 Nov 2026) |
| Tool calls on OpenRouter in a day | Not checked yet | 51.9K (for 10 Nov 2026) |
| Gpqa diamond | 64.1% (for 20 Jan 2025) | Not checked yet |
| Balrog | 19.5% (for 20 Jan 2025) | Not checked yet |
| Otis mock aime 2024 2025 | 55.6% (saturated benchmark: the top models all score near the ceiling) (for 20 Jan 2025) | Not checked yet |
| Artificial Analysis Intelligence Index | Not checked yet | 14.9 |
| LMArena text rating (Elo) | Not checked yet | 1,351 |
| Design Arena rating (Elo) | Not checked yet | 1,202 |
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.