Tools
Compare models
Add up to four models via ?models=slug-a,slug-b — or pick from any model page.
| Spec | Moonshot AI | Z.AI |
|---|---|---|
| Params | 1.1T / 32B | 753B |
| Architecture | moe | moe |
| Context | 262K | 1M |
| Weights | ||
| License | Custom licence | MIT License |
| 667.5 GBest | 475 GBest | |
| Cheapest hosted (out/1M) | $2.48 · OpenRouter | $2.42 · OpenRouter |
| Released | Apr 14, 2026 | Jun 16, 2026 |
| LiveBench | 70.5via Thinking | 73.2 |
| LiveBench Agentic Coding | 46.9via Thinking | 51.8 |
| LiveBench Coding | 78.6via Thinking | 79.7 |
| LiveBench Data Analysis | 65.1via Thinking | 73.7 |
| LiveBench Instruction Following | 64.4via Thinking | 62.3 |
| LiveBench Language | 75.1via Thinking | 76.2 |
| LiveBench Mathematics | 84.3via Thinking | 89.8 |
| LiveBench Reasoning | 79.4via Thinking | 78.6 |
| Arena Agent (IPS) | −0.011 | 0.071via Max |
| Arena Coding | 1514.8 | 1505.7via Max |
| Arena Creative Writing | 1429.5 | 1444.6via Max |
| Arena Hard Prompts | 1484.8 | 1488via Max |
| Arena Instruction Following | 1454.2 | 1464via Max |
| Arena Maths | 1479.3 | 1475.1via Max |
| Arena Text (overall) | 1460.7 | 1469.5via Max |
| Arena Code (WebDev) | 1509.5 | 1586.3via Max |
