Compare
gpt-5
Comparing 1 model
| GPT-5 | |
|---|---|
| Overview | |
| Status | Active |
| Released | 8/7/2025 |
| Summary | OpenAI's frontier flagship model with long context and advanced reasoning. |
| Pricing (per 1M tokens) | |
| Input | $1.25/1M |
| Output | $10/1M |
| Context and limits | |
| Context window | 272K |
| Max output | 128K |
| Modalities | |
| Supported | text, image |
| Capabilities | |
| Tool calling | Yes |
| Structured outputs | Yes |
| Open source | No |
| License | — |
| Benchmarks | |
| AIME 2024 % accuracy | 60 Self-reported |
| GPQA Diamond % accuracy | 72 Third-party |
| GSM8K % accuracy | 96.1 Self-reported |
| HumanEval pass@1 % | 94 Self-reported |
| MATH % accuracy | 87.1 Self-reported |
| MMLU % accuracy | 88 Self-reported |
| MMLU-Pro % accuracy | 76 Third-party |
| SWE-bench Verified % resolved | 48 Third-party |