
Sora 2
OpenAI's video model with synchronized audio and real-world physics — plus Cameos that put you in the scene.
About
Sora 2 (September 30, 2025) is OpenAI's second-generation video model: synchronized dialogue, sound effects, and music generated alongside video, physically-plausible motion (basketballs rebound correctly; gymnastics hold together), and multi-shot prompting. OpenAI called it the "GPT-3.5 moment for video."
The headline product feature is Cameos: verify your likeness once, then insert yourself (or consenting friends) into generated scenes with matched appearance and voice. Launched with a TikTok-style iOS social app (invite-only, US/Canada at launch; 1M+ downloads in under five days, per media reports). Closed and commercial — API access via OpenAI; free tier with "generous limits" constrained by compute. All outputs watermarked with metadata identifying AI generation.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
No standardized public benchmarks. Provider-claimed (OpenAI): physics accuracy and motion quality leadership — treat as provider-claimed. Independent: massive launch reception; comparative quality rankings vary by evaluator.
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 8/10
Scored 8/10 (verified tier). The reference commercial video model: synchronized audio+video, genuinely improved physics, and enormous real-world adoption (1M+ app downloads in days). Docked two points: (1) closed and commercial with compute-constrained free access; (2) quality claims are provider-reported, and the Cameo likeness system raises consent/moderation questions still being worked out.
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 10.0/10
- Maintenance · recency, releases, issue load
- 5.0/10
- Openness · license + docs
- 4.5/10
- Manual review · sanity read + red-flag check
- 9.0/10
- · provider-tier estimate — thinner evidence (no public adoption metric)
- · 1M+ app downloads in under five days (media-reported); no public API adoption metric
- · released 378 days ago
- · manual review: reference commercial video model; synchronized audio+video, improved physics; Cameo likeness system raises consent/moderation questions; license: proprietary — API/app only, no weights; compute-constrained free tier