Side-by-side comparison of AI model specs, benchmarks, and pricing.
| Spec | o3 OpenAI | Gemini 3.1 Pro Google |
|---|---|---|
| Tier | frontier | frontier |
| Release Date | 2025-04-16 | 2026-02-19 |
| Context Window | 200K | 1049K |
| Max Output | 100K | 66K |
| Input Price (/1M) | $2 | $2 |
| Output Price (/1M) | $8 | $12 |
| Arena Elo | 1,433 | 1,500 |
| MMLU | 89.4% | 92.6% |
| GPQA | 79.7% | 94.3% |
| MATH | 91.6% | 96.8% |
| HumanEval | 92.8% | 94.6% |
| SWE-bench | 71.7% | 80.6% |
| AIME | 96.7% | 91.2% |
| SimpleQA | 47.9% | 79.6% |
| Capabilities | vision, tool-use, code, reasoning | vision, tool-use, code, reasoning, agentic, audio |