OpenAI API Pricing (November 2026): GPT-6 Per-Token Prices, Batch, Flex, Fast Mode, Caching and a Cost Calculator

Last updated: 13 November 2026. All prices checked on the vendors’ official pages that day.

OpenAI API pricing is charged per token: you pay one rate for the tokens you send (input), a much lower rate for input that is served from the prompt cache, and a higher rate for the tokens the model writes back (output). In November 2026 the current flagship family is GPT-6, with four price points: gpt-6-astra, gpt-6.1-sol, gpt-6.1-terra and gpt-6-luna. This guide lists the official per-token prices, explains how Batch, Flex, Fast mode and caching change the bill, and gives you a calculator to estimate your own monthly cost.

Every number below was copied from OpenAI’s official API pricing page on 13 November 2026. If you are comparing vendors, read our LLM API pricing comparison. If your company needs help opening its own OpenAI account or connecting several models through one API, see our Overseas AI Token service: account setup + API access (PandaWM).

OpenAI API prices per 1M tokens (Standard tier)

A few things the table does not show:

• Short vs long context. The prices above apply to prompts up to 272K input tokens. Above that, GPT-6 models charge a long-context rate: for example gpt-6.1-sol goes from $2.00 / $10.00 to $4.00 / $15.00 per 1M input/output tokens, and gpt-6-astra from $10 / $50 to $20 / $75.
• Cache writes. GPT-6 models also price writing a prefix into the cache: on gpt-6.1-sol a 30-minute cache write is $2.50 per 1M tokens and a 12-hour write is $5.00. Reading from the cache is the cheap part ($0.10).
• Older models are still on the price list. gpt-5.4 is $2.50 / $15.00, gpt-5 is $1.25 / $10.00, gpt-5-nano is $0.05 / $0.40, gpt-4.1 is $2.00 / $8.00 and gpt-4o-mini is $0.15 / $0.60 per 1M input/output tokens.
• Promotions end. OpenAI notes that GPT-5.6 Sol’s promotional pricing runs until 30 November 2026, so check the page before you lock a budget around it.

Batch, Flex, Fast and Ultrafast: same model, different price

OpenAI sells the same model at several speeds. Standard is the default. Batch (asynchronous jobs) and Flex both cost half of Standard. Fast mode, which OpenAI renamed from Priority processing on 30 July 2026, costs twice Standard, and Ultrafast (available on gpt-6-astra and gpt-6.1-sol) costs six times Standard.

In practice most teams run user-facing chat on Standard, push evaluations, tagging, translation backlogs and report generation through the Batch API, and reserve Fast mode for a few latency-critical screens.

Other charges to include in a budget

• Web search tool: $10.00 per 1,000 calls, plus the retrieved search content billed as input tokens at the model’s rate.
• File search: storage at $0.10 per GB per day (1 GB free) and $2.50 per 1,000 tool calls.
• Hosted Shell / Code Interpreter containers: from $0.03 (1 GB) to $1.92 (64 GB) per 20-minute session.
• Regional processing (data residency): a 10% uplift for models released on or after 5 March 2026; FedRAMP endpoints also carry a 10% uplift.
• chat-latest (the model used in ChatGPT, exposed in the API) is listed at $5.00 input, $0.50 cached input and $30.00 output per 1M tokens. We cover it in
• ChatGPT API pricing vs ChatGPT plans
• .
• Fine-tuning is closed to new users; existing users can keep training on gpt-4.1 and gpt-4o models for now.

OpenAI API pricing calculator

Enter your expected monthly volume. If you do not know your token counts yet, read what a token is and how to estimate it first. The calculator uses Standard short-context list prices and optional Batch/Flex pricing.

Three worked examples

1. Customer-support assistant on gpt-6.1-terra: 50M input and 10M output tokens a month. Without caching: 50 × $1.00 + 10 × $5.00 = $100. If 30% of input is a repeated, cached system prompt: 35 × $1.00 + 15 × $0.05 + 50 = about $85.75.
2. Nightly product-catalogue tagging on gpt-6-luna via Batch: 200M input and 20M output tokens. 200 × $0.05 + 20 × $0.25 = $15 a month.
3. Agentic coding on gpt-6.1-sol: 300M input (60% cached) and 40M output. 120 × $2.00 + 180 × $0.10 + 40 × $10.00 = $658. The same workload on gpt-6-astra would be about five times that.

Output tokens usually dominate the bill on reasoning-heavy work, because reasoning tokens are billed as output. Watch output volume first.

Six ways to cut an OpenAI API bill

1. Start with the smallest model that passes your evaluation set, then move up only where quality fails.
2. Keep the start of every prompt identical (system prompt, tool definitions, examples) so it can be served from cache.
3. Move anything that does not need an instant answer to Batch or Flex.
4. Limit output length and ask for compact formats such as JSON with short keys.
5. Avoid crossing 272K input tokens unless you really need it; retrieval is usually cheaper than stuffing context.
6. Use separate projects with spend alerts for each product or customer, so one runaway job cannot burn the whole budget (see
7. rate limits
8. ).

How companies pay for the OpenAI API

API usage is billed to an OpenAI API organization, separately from any ChatGPT subscription. New accounts typically add prepaid credits on the billing page (quickstart). The API is only available in OpenAI’s supported countries and territories, and the account, payment method and billing details must belong to a real business that is entitled to use it. For a step-by-step setup, see how to get an OpenAI API key.

PandaWM helps companies set up and manage their own OpenAI organization (or that of a legitimately established overseas entity), with projects, keys, budgets and usage reporting, or connect to GPT, Claude and Gemini through one compliant unified API. Details: Overseas AI Token service: account setup + API access (PandaWM).

FAQ

How much does the OpenAI API cost per 1,000 tokens?

Divide the per-million price by 1,000. gpt-6.1-terra input is $1.00 per 1M tokens, so $0.001 per 1,000 input tokens; its output is $0.005 per 1,000 tokens.

What is the cheapest OpenAI model in the API?

Among current GPT-6 models, gpt-6-luna at $0.10 input and $0.50 output per 1M tokens, or half that through Batch/Flex. Older gpt-5-nano is $0.05 / $0.40.

Is the OpenAI API free?

There is no permanent free tier for production use. Eligible ChatGPT Pro and Business subscribers can claim monthly API credits, explained in our ChatGPT API pricing guide.

Does Batch pricing apply to every model?

Batch and Flex are listed at half of Standard for the GPT-6 family and most current models. Check the Batch table on the official page for the model you use.

Why is my bill higher than the calculator?

The calculator excludes long-context rates, cache writes, tools such as web search, regional-processing uplifts and tax. Reasoning tokens also count as output.

Is PandaWM an OpenAI reseller?

No. We help companies open and manage their own accounts or connect through a compliant unified API. Prices here are OpenAI’s public list prices.

Compliance note: Each model vendor only serves customers in its supported countries and regions and requires accurate account, identity and billing information. PandaWM helps companies use AI models through their own accounts or overseas entities that are legitimately established in supported regions, or through a compliant unified API. We do not sell, rent, share or transfer accounts or API keys, and we do not help anyone bypass regional restrictions, KYC, sanctions or payment controls. We are not an official partner or reseller of OpenAI, Anthropic, Google or any other vendor. Prices on this page are the vendors’ public list prices, not PandaWM prices.

Related guides

• Overseas AI Token service: account setup + API access (PandaWM)
• ChatGPT API pricing vs ChatGPT plans
• Gemini API pricing, free tier and billing tiers
• Claude API pricing: Opus, Sonnet, Haiku
• LLM API pricing comparison: cheapest providers
• How to get an OpenAI API key
• What is a token in AI? Token cost explained
• 简体中文：ChatGPT API 怎么购买（企业开通）
• 简体中文：大模型 API 价格对比

Official sources

• OpenAI API pricing
• OpenAI: Batch API
• OpenAI API quickstart
• OpenAI: supported countries and territories
• OpenAI: rate limits

All official pages checked on 2026-11-13. Prices are in USD per 1 million tokens (MTok), before tax. Vendors change prices; the official pages prevail.