Side-by-side comparison of AI model specs, benchmarks, and pricing.
| Spec | Claude Sonnet 4.5 Anthropic | GPT-5.2 OpenAI |
|---|---|---|
| Tier | frontier | frontier |
| Release Date | 2025-09-29 | 2025-12-11 |
| Context Window | 200K | 400K |
| Max Output | 64K | 128K |
| Input Price (/1M) | $3 | $1.75 |
| Output Price (/1M) | $15 | $14 |
| Arena Elo | 1,450 | 1,438 |
| MMLU | 88.7% | 92.8% |
| GPQA | 83.4% | 90.3% |
| MATH | 87% | 98% |
| HumanEval | 95% | 97% |
| SWE-bench | 55.8% | 68% |
| Capabilities | vision, tool-use, code, extended-thinking | vision, tool-use, code, reasoning, agentic |