Sora 2 vs Veo 3.1 — OpenAI and Google video compared
Sora 2 (OpenAI) and Veo 3.1 (Google DeepMind) both generate short clips with synchronised audio. Sora 2 allows longer durations at 720p; Veo 3.1 tops out short…
By LiliDi Editorial
Sora 2 (OpenAI) and Veo 3.1 (Google DeepMind) both generate short clips with synchronised audio. Sora 2 allows longer durations at 720p; Veo 3.1 tops out shorter but offers a 1080p tier and a cheaper Fast variant. Pick Sora 2: You need a longer single take (up to 12 seconds) or scene continuity across a dialogue beat. Pick Veo 3.1: You need 1080p output, or you want to iterate cheaply on Veo 3.1 Fast before spending on a flagship render. Verdict: Use Veo 3.1 Fast for exploration, Veo 3.1 for a 1080p final, and Sora 2 when the shot has to run longer than eight seconds. All three sit in the same picker on one credit balance. Provider: Sora 2 — OpenAI; Veo 3.1 — Google DeepMind. Durations in Lilidi: Sora 2 — 4s / 8s / 12s; Veo 3.1 — 4s / 6s / 8s. Quality tiers: Sora 2 — 720p; Veo 3.1 — 720p / 1080p. Aspect ratios: Sora 2 — 16:9, 9:16; Veo 3.1 — 16:9, 9:16. Native audio: Sora 2 — Yes —
synced dialogue and ambience; Veo 3.1 — Yes — dialogue, ambience and SFX. Credits per second: Sora 2 — 84; Veo 3.1 — 200 (Veo 3.1 Fast: 50). Typical use: Sora 2 — Longer story beats, dialogue scenes; Veo 3.1 — Cinematic hero shots; Fast variant for daily drafts. Q: Is Veo 3.1 Fast the same model? A: It is a faster, lower cost variant from the same family. It is a separate entry in the model picker at 50 credits per second. Q: Can I get 1080p from Sora 2? A: Sora 2 is offered at 720p in Lilidi. For 1080p output, use Veo 3.1 or upscale the result with the Upscale tool. Q: Which handles dialogue better? A: Both generate synced dialogue. Results depend heavily on prompt phrasing; test the same prompt on both before choosing.