Tools
Compare models
Add up to four models via ?models=slug-a,slug-b — or pick from any model page.
| Spec | Moonshot AI |
|---|---|
| Params | 2.8T / 104B |
| Architecture | moe |
| Context | 1M |
| License | |
| License terms | Custom licence |
| 1752.7 GBest | |
| Cheapest hosted (out/1M) | $12.75 · Makora |
| Released | Jun 13, 2026 |
| LiveBench | 79.19 |
| LiveBench Agentic Coding | 62.17 |
| LiveBench Coding | 81.45 |
| LiveBench Data Analysis | 78.73 |
| LiveBench Instruction Following | 71.36 |
| LiveBench Language | 85.53 |
| LiveBench Mathematics | 84.44 |
| LiveBench Reasoning | 90.67 |
| Arena Agent | 10th of 55 |
| Arena Agent · Recoverygets back on track after a command fails | 20th of 55 |
| Arena Agent · Steerabilitydoes what it was asked, and changes course when told | 25th of 55 |
| Arena Agent · Task outcomefinishes what the session set out to do | 7th of 55 |
| Arena Agent · Tool usereaches for the right one, and does not invent one | 9th of 55 |
| Arena Coding | 1541 |
| Arena Creative Writing | 1458 |
| Arena Hard Prompts | 1517 |
| Arena Instruction Following | 1486 |
| Arena Maths | 1499 |
| Arena Text (overall) | 1488 |
| Arena Code (WebDev) | 1660 |
