Tools
Compare models
Add up to four models via ?models=slug-a,slug-b — or pick from any model page.
| Spec | DeepSeek |
|---|---|
| Params | 1.6T / 37B |
| Architecture | moe |
| Context | 1M |
| License | |
| License terms | MIT License |
| 1008 GBest | |
| Cheapest hosted (out/1M) | $1.57 · OpenRouter |
| Released | Apr 22, 2026 |
| LiveBench | 71.57 |
| LiveBench Agentic Coding | 42.63 |
| LiveBench Coding | 69.99 |
| LiveBench Data Analysis | 74.54 |
| LiveBench Instruction Following | 62.35 |
| LiveBench Language | 78.13 |
| LiveBench Mathematics | 90.68 |
| LiveBench Reasoning | 82.69 |
| Arena Agent | 31st of 55 |
| Arena Agent · Recoverygets back on track after a command fails | 25th of 55 |
| Arena Agent · Steerabilitydoes what it was asked, and changes course when told | 33rd of 55 |
| Arena Agent · Task outcomefinishes what the session set out to do | 33rd of 55 |
| Arena Agent · Tool usereaches for the right one, and does not invent one | 40th of 55 |
| Arena Coding | 1506 |
| Arena Creative Writing | 1446 |
| Arena Hard Prompts | 1483 |
| Arena Instruction Following | 1461 |
| Arena Maths | 1457 |
| Arena Text (overall) | 1464 |
| Arena Code (WebDev) | 1582 |
