Openai Apis Plans Pricing
OpenAI prices the API per-model with usage-based, per-token rates (separate input and output prices), plus per-image, per-second, and per-request line items for non-text modalities. Account access is gated by usage tiers (Free, Tier 1 through Tier 5) which raise rate limits as cumulative spend grows; the tiers do not change unit prices. Batch API offers a 50% discount on most models. Enterprise customers can negotiate dedicated capacity (Provisioned Throughput) and custom contract pricing.
Plans
Pay-as-you-go per-token pricing for OpenAI's flagship GPT-5 family. Prices are per 1M tokens.
- 1M+ context window (select models)
- Streaming and structured outputs
- Function/tool calling
- Prompt caching discounts
Per-token pricing for reasoning, coding, deep research, and computer-use specialists.
- Specialized for coding, agents, browsing
- Tool-use optimized
Per-asset and per-second pricing for image generation, realtime audio, and video.
- Image generation and editing
- Realtime speech-to-speech
- Sora-2 video generation
Per-call and per-GB fees for built-in tools and storage attached to API usage.
- Built-in web search tool
- Files / Vector Stores
Asynchronous batch processing at approximately 50% off standard per-token rates with a 24-hour SLA.
- 24-hour completion window
- Higher throughput, lower cost
Provisioned Throughput Units (PTU), enterprise contracts, dedicated capacity, and custom data-handling terms negotiated with sales.
- Provisioned Throughput
- Enterprise data controls (zero retention)
- SOC 2 / HIPAA BAA available
- Dedicated support and SLAs