GPT-5 nano
Released · openai.com
Price
| Per 1M tokens | Input | Cached | Output |
|---|---|---|---|
| List price | $0.05 | $0.005 | $0.40 |
Price history
| Date | Input | Cached | Output |
|---|---|---|---|
| openai.com · 2 · Corroborated | $0.05 | $0.005 | $0.40 |
Context window
| Tokens | Context window | Input limit | Output limit |
|---|---|---|---|
| Default | 400,000 | 272,000 | 128,000 |
Context history
| Date | Context window | Input limit | Output limit |
|---|---|---|---|
| openai.com · 2 · Corroborated | 400,000 | 272,000 | 128,000 |
Benchmark results
| Benchmark | Score | Source verified |
|---|---|---|
| Humanity's Last ExamNo tools · high reasoning effort · reported later, in Introducing GPT-5 for developers | 8.7% | |
| SWE-bench Verifiedhigh reasoning effort; 477 of the 500 tasks · reported later, in Introducing GPT-5 for developers | 54.7% | |
| GPQA Diamondhigh reasoning effort; no tools · reported later, in Introducing GPT-5 for developers | 71.2% | |
| AIME 2025No tools · high reasoning effort · reported later, in Introducing GPT-5 for developers | 85.2% | |
| HMMT February 2025No tools · high reasoning effort · reported later, in Introducing GPT-5 for developers | 75.6% | |
| τ²-benchTelecom · high reasoning effort · reported later, in Introducing GPT-5 for developers | 35.5% | |
| τ²-benchRetail · high reasoning effort · reported later, in Introducing GPT-5 for developers | 62.3% | |
| τ²-benchAirline · high reasoning effort · reported later, in Introducing GPT-5 for developers | 41.0% |