Side-by-side comparison of AI model specs, benchmarks, and pricing.
| Spec | GPT-4.1 OpenAI | GPT-5.3 Codex OpenAI |
|---|---|---|
| Tier | mid | frontier |
| Release Date | 2025-04-14 | 2026-02-05 |
| Context Window | 1048K | 400K |
| Max Output | 33K | 128K |
| Input Price (/1M) | $2 | $1.75 |
| Output Price (/1M) | $8 | $14 |
| Arena Elo | 1,413 | 1,480 |
| MMLU | 86% | 93% |
| GPQA | 66.3% | 81% |
| MATH | 73.8% | 96% |
| HumanEval | 92.4% | 93% |
| SWE-bench | 54.6% | 80% |
| AIME | 52% | 94% |
| SimpleQA | 42.8% | 58% |
| Capabilities | vision, tool-use, code | vision, tool-use, code, reasoning, agentic |