Skip to content
Model infoModelsSnapshot: 15 Oct 2026

OpenAI / Sora 2

Sora 2

Retired OpenAI video model that generated clips with synchronized audio from text and image references. Historical rates and formats remain available for comparison.

Video generation
retiredClosed weightsVideo generation
Official documentation (opens in a new tab)Reviewed 14 Oct 2026
10 of 23 specification fields collected
Output resolution720×1280, 1280×720See specification notes below
Clip durationUp to 20 seconds per generationSee specification notes below
Input billingSee ratesOpenAI API
Output billingSee ratesPublished units and conditions below

Specifications

Provider-documented facts. “Not collected” means we have not verified that field yet, not that the feature is unsupported.

Context / input limit
Not collected

Not established by the reviewed model-specific sources.

Maximum output
Not collected

Not established by the reviewed model-specific sources.

Context caching
Not collected

Not established by the reviewed model-specific sources.

Reasoning
Not collected

Not established by the reviewed model-specific sources.

Tool calling
Not collected

Not established by the reviewed model-specific sources.

Structured output
Not collected

Not established by the reviewed model-specific sources.

Streaming
Not collected

Not established by the reviewed model-specific sources.

Batch processing
SupportedSource for Batch processing (opens in a new tab)

Historical POST /v1/videos Batch support added 2026-03-12. This is not current API availability.

Fine-tuning
Not collected

Not established by the reviewed model-specific sources.

Parameters
Not collected

Not established by the reviewed model-specific sources.

Architecture
Not collected

Not established by the reviewed model-specific sources.

Knowledge cutoff
Not collected

Not established by the reviewed model-specific sources.

Languages
Not collected

Not established by the reviewed model-specific sources.

Release date
Not collected

Not established by the reviewed model-specific sources.

Input modalities
Text, ImageSource for Input modalities (opens in a new tab)
Output modalities
Video, AudioSource for Output modalities (opens in a new tab)
Endpoint retirement date
2026-09-24Source for Endpoint retirement date (opens in a new tab)
Supported API routes
POST /v1/videosSource for Supported API routes (opens in a new tab)

Historical route, shut down with the Videos API.

Output duration
Up to 20 seconds per generationSource for Output duration (opens in a new tab)

Historical rolling-alias capability. Extensions had a separate workflow.

Output formats
MP4Source for Output formats (opens in a new tab)
Input file formats
JPEG, PNG, WebPSource for Input file formats (opens in a new tab)

Image references had to match the requested output resolution.

Output resolutions
720×1280, 1280×720Source for Output resolutions (opens in a new tab)
Snapshots and aliases
sora-2-2025-10-06, sora-2-2025-12-08Source for Snapshots and aliases (opens in a new tab)

Access & pricing

Prices belong to an endpoint and billing tier. A provider listing does not guarantee access for every account or region.

OpenAI API

Provider source (opens in a new tab)
API model ID
sora-2
Access
Retired: the Videos API and this model selector shut down on 2026-09-24. Documentation is retained for historical reference; no live access or replacement is implied.
Regions
Historical deployment regions not established for this model.
Endpoint retirement date
2026-09-24Source for Endpoint retirement date (opens in a new tab)
Availability
Retired native APISource for Availability (opens in a new tab)
Processing workflow
Asynchronous render job; poll status or use webhooks, then download MP4 output.Source for Processing workflow (opens in a new tab)

Job progress is not token streaming.

Reference images
JPEG, PNG or WebP matching output dimensions; multipart upload or JSON file_id/image_url reference.Source for Reference images (opens in a new tab)
Batch processing
SupportedSource for Batch processing (opens in a new tab)

Historical POST /v1/videos Batch support added 2026-03-12. This is not current API availability.

operations
Generation, Character references, Extensions, Video editsSource for operations (opens in a new tab)

Historical workflow operations; not independent checkpoints.

usageRestrictions
Historical guide rejects human-face input images and generation of real people. Content and reference restrictions applied.Source for usageRestrictions (opens in a new tab)

Prices retain their published currencies and billing units below.

Published billing units · original currency · observed 14 Oct 2026
CategoryUsageRateUnit
Historical video generation720p video output1$0.10generated video second

Rate conditions

  1. Historical standard rate retained on the model card for 720×1280 or 1280×720. Native API retired 2026-09-24. Not a current offer; no Batch discount, host credit conversion, or dated-snapshot tariff is inferred.
Pricing source (opens in a new tab)

Data & enterprise terms

Consult the provider API data controls for historical endpoint and account conditions; no Sora-specific zero-retention or current availability guarantee is inferred.

Read the service policy (opens in a new tab)

Caching

Caching can reuse earlier input processing. Availability, lifetime and billing depend on the endpoint. A cache hit still generates a new response.

01 / INITIAL INPUTProcess & retain

The system processes eligible input and retains reusable state.

02 / LATER REQUESTCheck available state

Reuse depends on matching input, session state and the cache policy.

03 / CACHE HITReuse & generate

Reusing state can reduce repeated work. Any charges follow the billing rules below.

OpenAI API

sora-2
Not verified

No model-specific prompt-cache contract established. Saved media and asynchronous jobs are not evidence of prompt caching.

How it works
Not collected

No model-specific prompt-cache contract established. Saved media and asynchronous jobs are not evidence of prompt caching.

Minimum reusable input
Not collected

No model-specific prompt-cache contract established. Saved media and asynchronous jobs are not evidence of prompt caching.

Lifetime & refresh
Not collected

No model-specific prompt-cache contract established. Saved media and asynchronous jobs are not evidence of prompt caching.

How to enable it
Not collected

No model-specific prompt-cache contract established. Saved media and asynchronous jobs are not evidence of prompt caching.

Usage fields
Not collected

No model-specific prompt-cache contract established. Saved media and asynchronous jobs are not evidence of prompt caching.

Billing rules
Not collected

No model-specific prompt-cache contract established. Saved media and asynchronous jobs are not evidence of prompt caching.

Limits & data boundaries
Not collected

No model-specific prompt-cache contract established. Saved media and asynchronous jobs are not evidence of prompt caching.

No verified cache rate for this endpoint. An empty rate is not a zero charge.

Caching reviewed 14 Oct 2026

What the measurements say

No benchmark observations have been added for this model. We do not turn a vendor description into a performance score.

Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.

The record behind the model

Why it is included

Named proprietary OpenAI video-generation release with official alias and snapshot documentation. Hosted API retirement does not make its weights public.

Classification evidence (opens in a new tab)

Lifecycle dates

Deprecation announced24 Mar 2026Source (opens in a new tab)
Videos API shut down24 Sept 2026Source (opens in a new tab)

Sources reviewed