Side-by-side comparison of Phi-3 Medium (Microsoft) and GPT-4o (OpenAI) — benchmarks, pricing, and capabilities.
| Phi-3 Medium Microsoft | GPT-4o OpenAI | |
|---|---|---|
| Category | LLMs | LLMs |
| Specifications | ||
| Context Window | 128K | 128K |
| Pricing (per 1M tokens) | ||
| Input Cost | Free | $2.50 |
| Output Cost | Free | $10.00 |
| Performance | ||
| Overall Score | 78.2 | 92.5 |
| ARC-Challenge | — | 85.8 |
| BigBench Hard | 68.0 | 87.2 |
| Chatbot Arena ELO | — | 1108.0 |
| DROP | 70.0 | 88.1 |
| GSM8K | 78.0 | 95.3 |
| HumanEval | 70.0 | 92.0 |
| MATH | 48.0 | 76.6 |
| MMLU | 78.0 | 88.7 |
| TruthfulQA | 52.0 | 70.1 |
| WinoGrande | — | 80.2 |
| Community | ||
| User Rating | ★ 4.2 | ★ 4.7 |
| Reviews | 312 | 1240 |