GPT-5 Nano: Pricing, Context, and Benchmarks
GPT-5 Nano is OpenAI's GPT-5 model released on 2025-08-07. It runs on a 400K-token context window, priced at $0.05/M input and $0.40/M output. Cheapest GPT-5 tier. Use for routing, classification, and short-form generation.
Specifications
| Provider | OpenAI |
|---|---|
| Family | GPT-5 |
| Released | 2025-08-07 |
| Status | active |
| Context window | 400K tokens |
| Max output | 64K tokens |
| Modalities | text |
| Capabilities | function calling json mode streaming |
| Tokenizer | o200k_base |
| API endpoint | https://api.openai.com/v1/responses |
| Documentation | Official docs ↗ |
| Last updated | 2026-05-29 |
What this model is good for
Based on benchmark scores and capability fit, GPT-5 Nano ranks well for:
- Cheapest Model with Vision — rank #1. $0.05/M input. Vision is text-only at this tier, but for OCR-style tasks the nano is unbeatable.
- High-Throughput Inference — rank #3. Fastest GPT-5 tier from OpenAI directly.
Cost calculator
For 1,000 tokens of input and 1,000 tokens of output per call, this model costs $0.0005 per call. For typical usage:
| Workload | Tokens / call (in/out) | $/call | $/1K calls | $/1M calls |
|---|---|---|---|---|
| Short Q&A | 200 / 100 | $0.000050 | $0.05 | $50 |
| Standard chat | 1K / 500 | $0.000250 | $0.25 | $250 |
| RAG with retrieval | 4K / 500 | $0.000400 | $0.40 | $400 |
| Long doc summary | 20K / 1K | $0.001400 | $1.40 | $1,400 |
| Long context (100K input) | 100K / 2K | $0.005600 | $5.60 | $5,600 |
For more precise estimates including caching discounts, use the LLM API Cost Calculator.
Other OpenAI models
Common comparisons
- GPT-5 Nano vs Codestral Side-by-side specs, pricing, and benchmarks
- GPT-5 Nano vs DeepSeek-V3 Side-by-side specs, pricing, and benchmarks
- GPT-5 Nano vs Gemini 2.5 Flash Side-by-side specs, pricing, and benchmarks
- GPT-5 Nano vs Qwen3-235B Side-by-side specs, pricing, and benchmarks
API setup
curl https://api.openai.com/v1/responses \
-H "Authorization: Bearer $API_KEY" \
-H "Content-Type: application/json" \
-d '{"model": "gpt-5-nano", "messages": [{"role": "user", "content": "Hello"}]}'See the official documentation for the full request/response schema. For pre-flight token counting before paying, use the appropriate token counter.