DeepSeek V2.5 (1210): benchmark results
Provider: DeepSeek. Released 2024-12-10. Access: Open.
Aggregate Rating 1554 ± 31, rank #1658 of 10916 rated models, from 28 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| BigCodeBench | 48.6 | Pass@1 (%) | 93.6 |
| Aider Code Editing Leaderboard | 72.2 | Exercises completed correctly after one retry, pass_rate_2 ( | 89.2 |
| Chatbot Arena (Text - Creative Writing) | 1310 | Arena Score | 46.6 |
| Chatbot Arena (Text - Entertainment, Sports & Media) | 1302 | Arena Score | 44.4 |
| Chatbot Arena (Text - Writing, Literature & Language) | 1315 | Arena Score | 44.1 |
| Chatbot Arena (Text - Software & IT Services) | 1364 | Arena Score | 42.9 |
| Chatbot Arena (Text - Coding) | 1374 | Arena Score | 42.2 |
| Chatbot Arena (Text - Multi-Turn) | 1323 | Arena Score | 40.9 |
| Chatbot Arena (Text - Instruction Following) | 1314 | Arena Score | 40.7 |
| Chatbot Arena (Text - Non-English) | 1306 | Arena Score | 40.6 |
| Chatbot Arena (Text - Chinese) | 1355 | Arena Score | 40.4 |
| Chatbot Arena (Text - Longer Query) | 1337 | Arena Score | 39.3 |
Interactive version: theaggregate.ai/model?slug=deepseek-v2-5-1210 · How It Works · Data refreshed daily, snapshot 2026-11-08.