Z.ai: GLM 5.2 wins on more metrics (5 of 9), but the right pick depends on what you optimize for — see the breakdown below.
| Metric | Anthropic: Claude Opus 4.7 | Z.ai: GLM 5.2 |
|---|---|---|
| Intelligence Index | — | 50.7✓ |
| Coding Index | 53.1✓ | 50.7 |
| GPQA Diamond | 89% | 90%✓ |
| Design Arena Elo | 1503✓ | 1369 |
| Speed (tokens/sec) | 57✓ | 30 |
| Latency | 1.3s✓ | 1.4s |
| Input price /M | $5.00 | $1.40✓ |
| Output price /M | $25.00 | $4.40✓ |
| Context window | 1M | 1.0M✓ |
| Capabilities | ReasoningToolsJSONVision | ReasoningToolsJSON |