Is OpenAI Moderation API Free? Moderation API Pricing Compared

Is OpenAI moderation API free? Yes. OpenAI's help center answers the question directly: "the Moderation endpoint is free for OpenAI API users, and usage of this tool doesn't count towards your monthly usage limits"^[1]. The API pricing page lists omni-moderation-latest with the price "Free"^[2], and the moderation guide repeats that the endpoint "is free to use"^[3].

Free does not mean unlimited, though. You still need an OpenAI API account, your requests are capped by usage-tier rate limits, and the model only scores some categories on images. This guide covers those conditions first, then puts the moderation API pricing of eight other options in one table so you can see what the alternatives cost per text item and per image. All prices were retrieved in October 2026.

TL;DR

• OpenAI's moderation endpoint is free for OpenAI API users and does not count toward monthly usage limits^[1].
• The limit is throughput. On the Free usage tier, omni-moderation-latest allows 250 requests per minute, 5,000 requests per day and 10,000 tokens per minute^[4].
• Coverage has edges. It reads text and images but not audio, accepts images up to 20 MB, and applies only 6 of its 13 categories to images^[3].
• Paid APIs bill different units: Azure per 1,000-character text record, Amazon Comprehend per 100 characters with a 300-character minimum, Hive per request, Sightengine per operation^[5][6][7][8].
• Mistral Moderation 2 is also listed as free, for text^[9][10].
• reAPI's Content Moderation is paid: it runs omni-moderation-latest as an async task billed per moderation unit. The current rate is on the model page^[11].

Is OpenAI moderation API free? What OpenAI's docs say

Three OpenAI pages agree. The help center article "Is the Moderation endpoint free to use?" says yes, for OpenAI API users, and adds that moderation calls do not count toward your monthly usage limits^[1]. The API pricing page lists one moderation model, omni-moderation-latest, at "Free"^[2]. The model page describes moderation models as "free models designed to detect harmful content"^[4].

The older text-only model, text-moderation-latest, carries the same "free models" wording, but OpenAI calls it the previous generation and expects the omni-moderation-* models "to be the best default moving forward"^[12].

So there is no per-call price to plan for. What you plan for instead is how many calls you can make and what the model actually checks.

The limits that come with the free endpoint

Rate limits by usage tier

Rate limits for omni-moderation-latest depend on your organization's usage tier^[4]:

The Free tier requires only that you are in an allowed geography. Build starts at $5 in total credit purchases, Launch at $100 and Grow at $500^[13]. In practice that means moderation is free to call, but higher throughput comes from having bought credits for other OpenAI models. When you exceed a temporary rate limit the API returns a 429 error, sometimes with a Retry-After header that tells you how long to wait^[13].

The 5,000 requests per day cap is the number to check first. A request can carry an array of strings^[3], so batching comments into one call stretches that cap much further than one call per comment.

What the model checks

omni-moderation-latest returns 13 categories. Six of them (self-harm, self-harm/intent, self-harm/instructions, sexual, violence, violence/graphic) accept text and images. The other seven, including harassment, hate, illicit and sexual/minors, are text only. If you send an image with no text, those seven come back as 0^[3]. The model does not classify audio, and image files are limited to 20 MB^[3]. Our image moderation API guide covers the image side in detail.

Two more caveats from the same guide. OpenAI plans to keep upgrading the underlying model, so policies built on category_scores thresholds "may need recalibration over time"^[3]. And the endpoint is not a child-safety tool: OpenAI says not to send known or suspected CSAM to it^[3].

Moderation API pricing compared: text and images

The table lists published list prices and free allowances for the moderation APIs we could verify on each vendor's own pricing page. Regional prices are for the region noted.

We left Google's paid Text Moderation rate out on purpose. Its pricing page lists a figure for the 50,000 to 10,000,000 band under a header that reads "Price per 100-character unit", while the same page prices other features per 1,000 units, and we could not tell which basis applies^[16]. Check the page or your console before budgeting.

What a "unit" means at each vendor

Per-unit prices are only comparable once you know the unit. Here is how each vendor counts:

• Azure: a text record is up to 1,000 Unicode code points. A 7,500-character input counts as 8 records; a 500-character input counts as 1^[5].
• Amazon Comprehend: 1 unit is 100 characters, and every request is charged at least 3 units (300 characters)^[14]. Toxicity detection is English only and takes up to 10 strings of up to 1 KB per call^[17].
• Google Natural Language: Text Moderation rounds each request up to the nearest 100 Unicode characters, including whitespace and markup^[16].
• Google Vision and Amazon Rekognition: per image. In Vision, each feature applied to an image is a separate unit^[15]; in Rekognition, running several APIs on one image counts as several images^[6].
• Hive: per request, with the self-serve tier limited to 100 requests per day^[7].
• Sightengine: per operation. Running nudity detection on one image is 1 operation, and requests that end in an error are not counted^[8].
• reAPI: per moderation unit. One unit is 1,000 words of text, summed across the request with each text item counted as at least 50 words, or one image. Every request bills at least one unit^[11].

An illustrative month: 100,000 comments

Assume 100,000 comments of 500 characters each in one month. These figures are our own arithmetic from the list prices above, not quotes:

• OpenAI: $0. Spread evenly that is about 3,334 requests a day, under the Free tier's 5,000 per day^[4].
• Azure Standard: 100,000 records × $0.38 / 1,000 = $38.00^[5].
• Amazon Comprehend: 5 units per comment, 500,000 units × $0.0001 = $50.00^[14].
• Hive: 100,000 × $0.50 / 1,000 = $50.00, but the self-serve cap of 100 requests per day allows about 3,000 a month, so this volume needs a sales contract^[7].

For 10,000 images a month the same arithmetic gives $0 on OpenAI, $7.50 on Azure, $10.00 on Rekognition once its 12-month free tier ends, and $13.50 on Google Vision SafeSearch after the first 1,000 free units^[5][6][15].

Where reAPI fits if OpenAI's endpoint is free

On reAPI, the same model id, omni-moderation-latest, is offered as Content Moderation, and it is paid. It is a different product shape from OpenAI's endpoint, and it is not a drop-in replacement for code written against OpenAI's SDK:

• Async. POST /api/v1/moderations returns a task id; you poll GET /api/v1/tasks/{id} until the status is completed, and the verdict sits under output.moderation.results. Polling is not charged^[18].
• Text and images. Input is a string, an array of strings, or text and image_url blocks. Images must be public http(s) URLs; base64 and data: URIs are rejected^[18].
• Per-unit billing, as described above, with failed requests refunded automatically. One request takes up to 500 items and 30,000 words^[18].

If you only need moderation and already have an OpenAI API account, OpenAI's free endpoint is the cheaper choice. reAPI makes sense when moderation sits next to other models you already call through reAPI, with one key, one credit balance and one task contract.

FAQ

Does OpenAI have moderation?

Yes. OpenAI offers a moderation endpoint backed by omni-moderation-latest, which classifies text and images against 13 harm categories^[3]. You can call it on its own or request moderation scores alongside a generated response in the Responses API^[3].

Is omni moderation latest free?

Yes. OpenAI's pricing page lists omni-moderation-latest as Free^[2], and calls do not count toward your monthly usage limits^[1]. Requests are still subject to rate limits. The same model on reAPI is billed per moderation unit; see the model page.

Openai moderation api rate limit

It depends on your usage tier: 250 requests per minute and 5,000 per day on Free, 500 per minute on Build, 2,000 on Launch and 5,000 on Grow, with token-per-minute caps from 10,000 to 500,000^[4]. Your current limits are shown under Settings, Organization, Limits in the OpenAI dashboard^[13].

Openai moderation api too many requests

"Too many requests" is HTTP 429. OpenAI returns it when a request exceeds a temporary rate limit and may include a Retry-After header; wait at least that long, add a small random delay, and retry with exponential backoff^[13]. Sending arrays of strings instead of one call per item reduces how often you hit the per-minute cap^[3].

Content moderation api free

OpenAI's moderation endpoint and Mistral Moderation 2 are listed as free^[2][9]. Several paid APIs have monthly free allowances: Azure gives 5,000 text records and 5,000 images, Google Natural Language 50,000 Text Moderation units, Google Vision 1,000 units, Amazon Rekognition 1,000 images for 12 months, and Sightengine 2,000 operations^{[5][16][15][6][8]}.

Text moderation api free

For text, OpenAI's endpoint and Mistral Moderation 2 have no listed charge^[2][9]. Mistral's model classifies text across 11 categories, including PII and jailbreaking^[10]. Azure's free tier covers 5,000 text records a month and Google's covers 50,000 Text Moderation units^[5][16]. Amazon Comprehend's toxicity detection is not on its free-tier list^[14].

Hive moderation api pricing

Hive's self-serve prices are $3.00 per 1,000 requests for Visual Moderation and $0.50 per 1,000 requests for Text Moderation, each limited to 100 requests per day; higher limits go through sales. Audio moderation is $0.03 per minute, and new developer accounts get $50+ in credits after adding a payment method^[7].

Which AI API is completely free?

Among moderation APIs, OpenAI's moderation endpoint and Mistral Moderation 2 have no listed price^[2][9]. Both still require an account with that provider, and OpenAI's endpoint is capped by rate limits^[4].

Choosing between free and paid moderation

If the question is cost alone, the answer is short: OpenAI moderation API is free, and for text and images inside its 13 categories it is hard to beat on price. Plan around the 5,000 requests per day on the Free tier, batch your inputs, and handle 429 with backoff.

Paid moderation APIs earn their price on different grounds: Azure's severity levels and custom categories, Rekognition's video moderation and detailed image labels, Sightengine's 120+ image classes. Our best content moderation API comparison walks through those choices by use case. If you want omni-moderation-latest on the same key and task contract as your other reAPI models, the current rate is on the Content Moderation model page. For a pure cost decision, though, the answer to "is OpenAI moderation API free" is yes.

References

1. OpenAI. Is the Moderation endpoint free to use? Retrieved October 2026 from help.openai.com/en/articles/4936833-is-the-moderation-endpoint-free-to-use
2. OpenAI. API pricing. Retrieved October 2026 from developers.openai.com/api/docs/pricing
3. OpenAI. Moderation guide. Retrieved October 2026 from developers.openai.com/api/docs/guides/moderation
4. OpenAI. omni-moderation model page. Retrieved October 2026 from developers.openai.com/api/docs/models/omni-moderation-latest
5. Microsoft Azure. Content Safety pricing. Retrieved October 2026 from azure.microsoft.com/en-us/pricing/details/content-safety
6. Amazon Web Services. Amazon Rekognition pricing. Retrieved October 2026 from aws.amazon.com/rekognition/pricing
7. Hive. Pricing. Retrieved October 2026 from thehive.ai/pricing
8. Sightengine. Pricing. Retrieved October 2026 from sightengine.com/pricing
9. Mistral AI. Mistral Moderation 2 model card. Retrieved October 2026 from docs.mistral.ai/models/mistral-moderation-26-03
10. Mistral AI. Moderation and guardrailing. Retrieved October 2026 from docs.mistral.ai/studio/conversations/moderation
11. reAPI. Content Moderation model page. reapi.ai/models/content-moderation
12. OpenAI. text-moderation model page. Retrieved October 2026 from developers.openai.com/api/docs/models/text-moderation-latest
13. OpenAI. Rate limits. Retrieved October 2026 from developers.openai.com/api/docs/guides/rate-limits
14. Amazon Web Services. Amazon Comprehend pricing. Retrieved October 2026 from aws.amazon.com/comprehend/pricing
15. Google Cloud. Cloud Vision API pricing. Retrieved October 2026 from cloud.google.com/vision/pricing
16. Google Cloud. Cloud Natural Language pricing. Retrieved October 2026 from cloud.google.com/natural-language/pricing
17. Amazon Web Services. Trust and safety, Amazon Comprehend Developer Guide. Retrieved October 2026 from docs.aws.amazon.com/comprehend/latest/dg/trust-safety.html
18. reAPI. Content Moderation API docs. reapi.ai/docs/content-moderation