Side-by-side comparison of AI model specs, benchmarks, and pricing.
| Spec | GPT-5.3 Codex OpenAI | Claude Opus 4.6 Anthropic |
|---|---|---|
| Tier | frontier | frontier |
| Release Date | 2026-02-05 | 2026-02-05 |
| Context Window | 400K | 200K |
| Max Output | 128K | 128K |
| Input Price (/1M) | $1.75 | $5 |
| Output Price (/1M) | $14 | $25 |
| Arena Elo | 1,480 | 1,496 |
| MMLU | 93% | 92.5% |
| GPQA | 81% | 91.3% |
| MATH | 96% | 97.6% |
| HumanEval | 93% | 97% |
| SWE-bench | 80% | 72.5% |
| AIME | 94% | 83.3% |
| SimpleQA | 58% | 43.2% |
| Capabilities | vision, tool-use, code, reasoning, agentic | vision, tool-use, code, extended-thinking, agentic |