Side-by-side comparison of DeepSeek Coder V2 (DeepSeek) and Claude 3.5 Sonnet (Anthropic) — benchmarks, pricing, and capabilities.
| DeepSeek Coder V2 DeepSeek | Claude 3.5 Sonnet Anthropic | |
|---|---|---|
| Category | Code Models | LLMs |
| Specifications | ||
| Context Window | 128K | 200K |
| Pricing (per 1M tokens) | ||
| Input Cost | $0.14 | $3.00 |
| Output Cost | $0.28 | $15.00 |
| Performance | ||
| Overall Score | 84.5 | 91.2 |
| ARC-Challenge | 96.3 | 93.9 |
| BigBench Hard | — | 86.0 |
| Chatbot Arena ELO | — | 1242.0 |
| DROP | — | 86.5 |
| GSM8K | 65.1 | 94.1 |
| HumanEval | 90.8 | 90.2 |
| MATH | 59.8 | 75.2 |
| MBPP | 86.5 | — |
| MMLU | 89.2 | 74.8 |
| TruthfulQA | 72.4 | 77.4 |
| WinoGrande | 82.0 | 83.7 |
| Community | ||
| User Rating | ★ 4.5 | ★ 4.6 |
| Reviews | 445 | 890 |