← All models

Sora 2

Video Retired Full comparison ↗

openai/sora-2 · by OpenAI · Diffusion transformer (DiT)

Pricing — 1 offering(s)

OpenAI

View on OpenAI ↗

Standard 720p

Retired — no longer sold · last price $0.10 / second, until 2026-09-23

$0.10 / second Historical 2026-06-13 → 2026-09-23

Pro 720p

Retired — no longer sold · last price $0.30 / second, until 2026-09-23

$0.30 / second Historical 2026-06-13 → 2026-09-23

Pro 1024p

Retired — no longer sold · last price $0.50 / second, until 2026-09-23

$0.50 / second Historical 2026-06-13 → 2026-09-23

Pro 1080p

Retired — no longer sold · last price $0.70 / second, until 2026-09-23

$0.70 / second Historical 2026-06-13 → 2026-09-23

Showing the active price and any recorded history. Full pricing history is available via the paid API — see API docs.

See how Sora 2 fits into a cost-aware routing setup

Capability profile

How Sora 2 rates across core capability dimensions, with the task-level evidence behind each rating.

Prompt adherence Strong

Handles complex multi-subject prompts well; strong text encoder integration.

temporal consistency Strong

Spatiotemporal attention produces stable, coherent subjects across frames.

motion quality Strong

Physically plausible motion; notable improvement over Sora 1 on fluid and cloth.

native audio Weak

No native audio; outputs silent video.

camera control Moderate

Responds to camera-related prompt words but lacks explicit camera control parameters.

Resolution ceiling Moderate

Max 1080p (Pro tier); 720p for Standard. No 4K.

Inference speed Moderate

Generation time not publicly disclosed; typically several minutes per clip.

clip duration ceiling Moderate

Up to 20 seconds per call.

cost efficiency Weak

video editing Weak

T2V and I2V only via OpenAI API; no editing of existing clips.

Ratings are estimated — limited independent data is available for this model.

Operator guidance

Choose Sora 2 Standard (720p) for cost-sensitive workflows needing solid T2V quality. Choose Sora 2 Pro for higher resolution output or maximum quality. Strong all-rounder; outperform on complex compositional prompts. Not suited where audio is required (no native audio) or where budget is very tight (Pro 1080p is $0.70/s, among the most expensive in the market).

Use cases

Professional content production and video marketing
Cinematic storytelling with complex scenes
Product visualisation from text prompts
Research and creative prototyping

Limitations

Retired: OpenAI shut down the Videos API and all Sora 2 models on 2026-09-24, with no recommended replacement (developers.openai.com/api/docs/deprecations)
No native audio generation
Pro tier at 1080p ($0.70/s) is expensive for long clips
No explicit camera control parameters (prompt-driven only)
Maximum 20 seconds per generation call
Content policy restrictions limit certain creative uses

Citations

Sora: Video generation models as world simulators — Original Sora 1 technical report; Sora 2 details not separately published
OpenAI Sora API documentation