by OpenAILLM APIs
Per-token access to OpenAI's GPT-6 models and hosted tools through one API, paid from prepaid credits.
One API key for OpenAI's GPT-6 models and hosted tools, priced per token with published batch, flex and caching discounts.
The OpenAI API gives you OpenAI's current models (GPT-6 Astra, GPT-6.1 Sol, GPT-6 Luna) plus built-in tools such as web search, file search and code containers through one Responses API, with official SDKs in six languages. Every current model's token price is on one public page, and Batch and Flex cut it in half. Teams that want OpenAI's models and agent tooling without running inference get a pay-as-you-go service with a SOC 2 Type 2 report and an ISO 27001 certificate that cover the API.
Against 3 other products we reviewed in LLM APIs: what you pay for with each, the five access facts, the plan limits we recorded and each review's bottom line. The note under the table says how each one was checked.
| Metric | OpenAI API | Mistral AI API | Anthropic API (Claude) | Google Gemini API |
|---|---|---|---|---|
| What you're paying for | Saves engineering timeOne API key for OpenAI's GPT-6 models and hosted tools, priced per token with published batch, flex and caching discounts. | Worth the priceLow per-token prices on small open-weight models, with EU or US processing you can choose yourself and spending caps per workspace. | Worth the priceClaude models from Haiku 5.5 at $0.10 / $0.50 per 1M tokens up to Fable 5.1, with no training on API data and a BAA you can sign in the console. | Easy to adoptGoogle's Gemini models by the token in US dollars, with an ongoing free tier to start and batch, flex and caching discounts to scale. |
| Overall score | 7.1 | 7.4Best | 7.2 | 5.3 |
| Free option | –Not measured | Free plan | Free trial | Free plan |
| Card to start | –Not measured | No card needed | –Not measured | No card needed |
| Getting access | Self-serve | Self-serve | Self-serve | Self-serve |
| Starter priceCheapest paid plan | $0.10/1M input tokens | $0.10/1M input tokens | $0.10/1M input tokens | $0.25/1M input tokens |
| Pricing model | Usage-based | Usage-based | Usage-based | Usage-based |
| Pricing page | Public prices | Public prices | Public prices | Public prices |
| How you pay | Prepaid credits ($5 minimum), expire after 1 year | Prepaid pay-as-you-go credits, optional auto recharge | Prepaid credits by card, expire after 1 year; invoicing through sales | Prepaid credits ($5 minimum), expire after 12 months |
| Free option | No free allowance stated | Free mode, no card; allowance not stated publicly | One-time free credits for new users, amount not stated | Ongoing free tier on select models, no card; limits shown in AI Studio |
| Reference scenario (100M in + 20M out tokens/mo) | $20 on gpt-6-luna, computed | $12 on ministral-3b-2512, computed | $20 on claude-haiku-5-5, computed | $55 on gemini-3.1-flash-lite, computed |
| Batch, cache and other discounts | Batch and Flex 50% off; cached input 90% off | Batch 50% off; cached input 90% off | Batch 50% off; cache hits 90% off on Haiku 5.5 | Batch and Flex (preview) 50% off; cached input 90% off |
| Trained on API data | No, unless you opt in | Unless you opt out (Free mode and paid) | No (Commercial Terms) | Paid: no. Free tier: may be used to improve Google products |
| Default retention | Abuse logs up to 30 days | 30 rolling days for abuse monitoring | Deleted within 30 days | Abuse logs for a "limited period", days not stated |
| Zero data retention | By approval, through sales | By approval, pay-as-you-go, stateless endpoints only | By agreement, through sales | Not guaranteed: paid prompts are logged for abuse monitoring (docs point to Gemini Enterprise Agent Platform, formerly Vertex AI) |
| SOC 2 / ISO 27001 | Both, covering the API | Both, stated in the help center | Both, listed for the API | Not stated for this API |
| HIPAA BAA | Standard BAA accepted in the API Platform, if eligible | Not stated | Standard BAA signed in the Console, if eligible | Not stated for this API |
| Data regions | 10 storage regions by approval; processing in US, EU, UAE | EU and US endpoints at +10% | US-only inference at 1.1x; otherwise global | No region choice stated |
| Official SDKs | Python, TypeScript/JavaScript, Java (beta), Go (beta), C#, Ruby | Python, TypeScript | Python, TypeScript, C#, Go, Java, PHP, Ruby | Python, JavaScript/TypeScript, Go, Java, C# |
| Console SSO | Custom/Enterprise billing plan | SAML on the Enterprise plan | Through sales | Not stated (Google account sign-in) |
| Bottom line | From its public pages, the OpenAI API has a complete per-token price list, published batch, flex and caching discounts, and audit reports that cover the API. Stricter data controls, residency and SSO go through sales, and a few docs pages lag behind the price list. | From its public pages, the Mistral AI API has the lowest reference price in this comparison, EU or US endpoints you can pick yourself and spending caps. Training use is an opt-out, rate limits sit behind sign-in, and ZDR and SSO need approval or the Enterprise plan. | From its public pages, the Anthropic API has a complete per-token price list, discounts that combine and strong data terms, with a self-serve HIPAA BAA. Region choice is US only, zero data retention and SSO go through sales, and a few pages disagree on billing details. | From its public pages, the Gemini API pairs an ongoing free tier on current models with a detailed USD price list and published batch, flex and caching discounts. Certifications, zero data retention and data regions sit with Google Cloud's Agent Platform, and a few billing pages disagree. |
Evidence: OpenAI API: facts checked 2026-10-15 from its public sign-up and pricing pages, no completed hands-on run; Mistral AI API: facts checked 2026-10-15 from its public sign-up and pricing pages, no completed hands-on run; Anthropic API (Claude): facts checked 2026-10-15 from its public sign-up and pricing pages, no completed hands-on run; Google Gemini API: facts checked 2026-10-15 from its public sign-up and pricing pages, no completed hands-on run. Prices in USD as listed, excluding tax. "–" means not recorded.
Review from public pages
Written from OpenAI API's public pages (pricing, security or privacy, and docs) as read on Oct 15, 2026. We did not sign up, and no hands-on run is published yet.
We read OpenAI's API pricing page, the models page and the model pages for GPT-6 Luna and the older nano models, the deprecations page, the data controls, rate limit, spend limit, batch, flex, prompt caching, structured outputs and SDK guides, and the trust portal on Oct 15, 2026; we did not sign up or call the API. The openai.com marketing, policy and help pages returned an error to our fetches that day, so we read their Internet Archive copies from Jun 18 to Oct 12, 2026 (the API page, the security and privacy page, the DPA, the enterprise privacy page, and the prepaid billing, SSO and BAA help articles).
The price list is complete and easy to use: every GPT-6 and GPT-5.6 model has standard, batch, flex and fast prices per 1M tokens, plus cached-input and cache-write prices. Billing is prepaid: the help center says the minimum purchase is $5, auto-reload is on by default, and purchased credits expire after a year. Usage tiers (Build, Launch, Grow) only lift rate limits and the monthly cap.
Picking the scenario model took care. GPT-5 nano is the cheapest row, but its only snapshot is on the deprecations page, as are GPT-4.1 nano and GPT-4o mini, so the reference price uses GPT-6 Luna, the current low-cost model. A few pages need a question: the rate limit guide names a Free tier and the quickstart mentions a free test request, but no page states a free allowance; the production guide still speaks of prices per 1,000 tokens; and the pricing page lists models the deprecations page says were shut down in September.
Every GPT-6 and GPT-5.6 model lists standard, batch, flex and fast prices per 1M tokens, with cached-input and cache-write prices beside them. Batch and Flex are both 50% off.
vs Anthropic API (Claude): Anthropic API (Claude) also has a 50% batch discount and 0.1x cache hits, but no published flex-style lower rate.
The trust portal lists a SOC 2 Type 2 report (Jul 1, 2025 to Jun 30, 2026) and an ISO 27001 certificate for the API, and the DPA applies with the Services Agreement.
Compared: Google's compliance scope lists name Gemini Enterprise Agent Platform, not the Gemini Developer API.
Storage can be kept in 10 regions, including the US, Europe, the UK, Japan and India, and requests can be processed in the US, Europe or the UAE. Eligibility goes through sales.
vs Anthropic API (Claude): Anthropic API (Claude) offers only a US option, at 1.1x the price; Google Gemini API states no region choice.
Hard monthly spend limits and alerts can be set per organization and per project; at a hard limit, requests are refused with a 429 error.
Compared: DeepSeek API states no spend limit; spending is bounded only by the prepaid balance.
The libraries page lists official SDKs for Python, TypeScript/JavaScript, Java (beta), Go (beta), C# and Ruby, and Structured Outputs follow a supplied JSON Schema.
vs Mistral AI API: Mistral AI API's official SDKs are Python and TypeScript only.
The rate limit guide lists a Free tier ($100 a month usage limit, allowed countries only) and the quickstart mentions a free test request, but no page states a free token amount, models or rate limits.
Workaround: Budget the $5 minimum credit purchase for a first test.
New accounts buy credits first ($5 minimum, $10 default). Auto-reload is turned on during setup, and purchased credits expire after 1 year and are non-refundable except in listed cases.
Workaround: Turn auto-reload off or set a monthly reload limit, and add a hard spend limit per project.
Zero data retention needs OpenAI approval, and data residency outside the US also needs approved abuse-monitoring controls and a Modified Retention amendment. The standard HIPAA BAA can be accepted in the API Platform only once an organization has an established history of API usage.
Workaround: Set store to false on Responses requests you do not need kept; otherwise they are stored for at least 30 days.
The production guide quotes prices per 1,000 tokens, the pricing page lists models shut down on Sep 28, 2026, and a GPT-5.6 Sol promotion note gives no regular price.
Workaround: Use the pricing page and the deprecations page as the sources; GA models get at least 6 months of notice.
OpenAI bills per token for each model from prepaid credits: the minimum purchase is $5, auto-reload is on by default, and purchased credits expire after a year. Usage tiers (Build at $5 of purchases, Launch at $100, Grow at $500) raise rate limits and the monthly cap; they are not plans. Our scenario on GPT-6 Luna is 100 × $0.10 + 20 × $0.50 = $20 a month at standard rates; Batch or Flex halves it, and cached input costs $0.01 per 1M. Costs climb with the flagship (GPT-6 Astra at $10 / $50 would be $2,000 for the same tokens, our arithmetic), Fast mode at 2x, prompts over 272K tokens, hosted tools such as web search at $10 per 1,000 calls plus content tokens, and a 10% uplift on data residency endpoints. Taxes are not stated on the pages we read.
There is little to export: you own your inputs and outputs, Responses objects are stored for at least 30 days by default, and files can be deleted or set to expire. Assistants objects stay until you delete them. Fine-tuned models remain callable until their base model is retired; a weight download is not stated. Moving away mostly means changing the base URL and model names, since Gemini, Mistral, DeepSeek and Anthropic document endpoints in OpenAI's request format.
From its public pages, the OpenAI API has a complete per-token price list, published batch, flex and caching discounts, and audit reports that cover the API. Stricter data controls, residency and SSO go through sales, and a few docs pages lag behind the price list.
| Limit | As recorded |
|---|---|
| How you pay | Prepaid credits ($5 minimum), expire after 1 year |
| Free option | No free allowance stated |
| Reference scenario (100M in + 20M out tokens/mo) | $20 on gpt-6-luna, computed |
| Batch, cache and other discounts | Batch and Flex 50% off; cached input 90% off |
| Trained on API data | No, unless you opt in |
| Default retention | Abuse logs up to 30 days |
| Zero data retention | By approval, through sales |
| SOC 2 / ISO 27001 | Both, covering the API |
| HIPAA BAA | Standard BAA accepted in the API Platform, if eligible |
| Data regions | 10 storage regions by approval; processing in US, EU, UAE |
| Official SDKs | Python, TypeScript/JavaScript, Java (beta), Go (beta), C#, Ruby |
| Console SSO | Custom/Enterprise billing plan |
Read from the vendor's public pages on Oct 15, 2026. No hands-on run is published yet. How we check
We have not published a sign-up run for OpenAI API yet. The access facts above were read from its sign-up and pricing pages on Oct 15, 2026.
We have not published API calls for OpenAI API yet.
OpenAI API lists 4 plans. The cheapest paid plan, GPT-6 Luna, costs $0.10/1M input tokens.
Prices as listed, USD, excluding tax; observed 2026-10-15.
Source: developers.openai.com pricing page. Prices change; the observed date above is when our agent read them.
We read the pricing, model, data-usage, trust and docs pages of three LLM APIs on the same day and scored them on the same six indicators. What 100 million input and 20 million output tokens a month cost, how each one treats your data, and who should pick which.
