OpenAI · video model · Released Oct 6, 2025Sora 2
Sora 2 is a video model from OpenAI, released Oct 6, 2025. It handles text to video and image to video.
- Arena score
- 1,343#17 of 48 · text to video
- Starting price
- —Not published
- Longest clip
- —Not published
- Resolution
- —No audio track
How it rates and what it costs
- Text to video arena
- 1,343 ±6#17 of 48 · 71,947 votes
Arena scores come from blind, side-by-side human votes on the same prompts (higher is better; ± is the 95% range). Boards updated Sep 22, 2026.
The details
- Lab
- OpenAI
- Type
- Video model
- Released
- Oct 6, 2025
- Inputs
- Text prompts, Reference images
- Output
- Video
- Can do
- Text to video, Image to video
- Weights
- Proprietary
- Availability
- 1 API host
More video models from OpenAI
About Sora 2
How good is Sora 2?
In blind side-by-side votes, Sora 2 scores 1,343 for text to video (#17 of 48). Higher is better; scores within about 12 points of each other are effectively tied.
Who makes Sora 2?
Sora 2 is a video model from OpenAI, released Oct 6, 2025.
What can Sora 2 do?
Sora 2 accepts text prompts and reference images and produces video. In practice: text to video and image to video.
Is Sora 2 open source?
No. Sora 2 is proprietary; you use it through an API or partner apps.
Where can I use Sora 2?
It is listed by 1 API host we track.
What are the alternatives to Sora 2?
The closest video models from other labs, by release date, are Veo 3.1 (Google, Oct 15, 2025), Wan v2.5 Preview (Alibaba (Qwen), Sep 24, 2025), Kling v2.6 (Kling AI, Dec 3, 2025) and Seedance v1.0 Pro Fast (ByteDance Seed, Oct 24, 2025).