All models / OpenAI
GPT-5
OpenAI uses implicit prompt caching (no separate cache-write fee). Reasoning effort levels (minimal/low/medium/high) affect output token usage.
At a glance
- Provider
- OpenAI
- Family
- gpt-5
- Released
- 2025-08
- Status
- Active
- Context window
- 400K
- Max output
- 128K
- Knowledge cutoff
- 2024-09
- API id
- gpt-5
Pricing
USD per 1M tokens.
- Input
- $1.25
- Output
- $10.00
- Cache read
- $0.125
- Cache write (5m)
- —
- Cache write (1h)
- —
- Batch input
- $0.625
- Batch output
- $5.00
Capabilities
- Tool use
- Yes
- Structured outputs
- Yes
- Vision
- Yes
- Extended thinking
- Yes
- Prompt caching
- Yes
- Batch API
- Yes
- Fine-tuning
- Yes
- Open weights
- No
Modalities
- Input
- text, image
- Output
- text
Compare with
Links
Last verified 2026-04-26. License: proprietary.