| Spec | GPT-4o | o3 |
|---|
| Provider | OpenAI | OpenAI |
| Release Date | May 2024 | Apr 2025 |
| Knowledge Cutoff | Oct 2023 | 2024-05-31 |
| Parameters | unknown | Undisclosed |
| Context Window | 128K | — |
| Max Output | — | — |
| Open Source | No | No |
| License | proprietary | proprietary |
| Tokenizer | — | — |
| Modality | — | — |
| Reasoning Model | No | No |
| Moderated | No | No |
| Price (per 1M tokens) | GPT-4o | o3 |
|---|---|---|
| Input | $2.50 | — |
| Output | $10.00 | — |
| Feature | GPT-4o | o3 |
|---|---|---|
| Text Input | ||
| Image Input (Vision) | ||
| Audio Input | ||
| Video Input | ||
| File Input | ||
| Image Output | ||
| Audio Output | ||
| Tool Use | ||
| Structured Output (JSON) | ||
| Streaming | ||
| Reasoning Tokens | ||
| Web Search | ||
| Temperature Control | ||
| Top-P Sampling | ||
| Stop Sequences | ||
| Seed (Reproducibility) | ||
| Log Probabilities |
| Benchmark | GPT-4o | o3 |
|---|---|---|
| IFEval | 83.6 | — |
| MMLU-Pro | 74.5 | — |
| Chatbot Arena Elo | 1285 | — |
| Benchmark | GPT-4o | o3 |
|---|---|---|
| SWE-bench Verified | 33.2 | 69.1 |
| HumanEval | 90.2 | — |
| Benchmark | GPT-4o | o3 |
|---|---|---|
| GSM8K | 95.8 | — |
| MATH | 76.6 | — |
| AIME 2024 | 13.4 | — |
| AIME 2025 | — | 86.4 |
| FrontierMath | — | 15.8 |
| Benchmark | GPT-4o | o3 |
|---|---|---|
| BigBench-Hard | 87.3 | — |
| ARC-Challenge | 96.4 | — |
| GPQA Diamond | 53.6 | 83.3 |
| Humanity's Last Exam | — | 14.7 |
| ARC-AGI v2 | — | 6.5 |
| Benchmark | GPT-4o | o3 |
|---|---|---|
| WinoGrande | 85.7 | — |
| HellaSwag | 96.4 | — |
| Benchmark | GPT-4o | o3 |
|---|---|---|
| MMMU | 69.1 | 82.9 |
| MMMU-Pro | — | 76.4 |
| CharXiv Reasoning | — | 78.6 |
| Benchmark | GPT-4o | o3 |
|---|---|---|
| TruthfulQA | 73.5 | — |
| Benchmark | GPT-4o | o3 |
|---|---|---|
| BrowseComp | — | 49.7 |