Quick verdict
Seedance 2.5 takes the quality evidence: it outscores veo-3.1-audio on both LMArena boards — 1474[5] to 1364[5] on text-to-video, 1477[6] to 1398[6] on image-to-video — and generates 30 s[4] in one pass where Veo 3.1 stops at 8 s[1]. Veo 3.1 stays the pick for documented delivery: 4K output on 8-second clips[1], extensions to 148 s[1], and a public price for every tier starting at $0.05/s[3] per second — none of which ByteDance matches in print.
- Pricing Depends
- Output specs Seedance 2.5
- Features Tie
- Access Tie
- Public leaderboards Seedance 2.5
Category winners are computed from the sourced rows in the tables below; “Depends” means there was not enough comparable public data to call it.
Summary
Google brought Veo 3.1 to the Gemini API on 2025-10-15[7], where it remains a paid preview[7]; ByteDance Seed followed with Seedance 2.5 on 2026-07-31[4], and its API is now generally available on BytePlus ModelArk[8]. Each generates native audio with the picture[2][4], and each can lengthen an existing video rather than only starting fresh.
They solve length in opposite ways. Seedance 2.5 produces up to 30 s[4] in a single pass and appends further shots through multi-round extension with no published ceiling[4]. Veo 3.1 generates 8 s[1] at a time, then extends 7 seconds per round to 148 s[1] at 720p[1]. Reference input is lopsided: 50[4] mixed inputs per pass for Seedance against 3[1] images for Veo. Resolution is lopsided the other way — Veo renders 1080p and 4K on 8-second clips[1], while the only official Seedance 2.5 figures are 480p and 720p[8].
Both arena boards list both models, and Seedance 2.5 leads each: rank #7[5] against #12[5] on text-to-video and rank #4[6] against #13[6] on image-to-video, with margins outside the confidence intervals. Pricing transparency runs the other way entirely: Google publishes per-second rates for every tier, from $0.05/s[3] (Lite, 720p) to $0.60/s[3] (Standard, 4K)[3], while ByteDance's API bills per million tokens and states no first-party per-second figure.
Snapshot
| Spec | Veo 3.1 | Seedance 2.5 |
|---|---|---|
| Status | Paid preview in the Gemini API (Google's current model list names veo-3.1-generate-preview and veo-3.1-lite-generate-preview as Preview models; Veo 3.1 Fast rates are still published on the pricing page); also served via Vertex AI, the Gemini app and Flow[7] | Generally available — API fully available to clients on BytePlus ModelArk; live in Jimeng Web and Doubao Pro[8] |
| Developer | Google[7] | ByteDance Seed[4] |
| Released | 2025-10-15[7] | 2026-07-31[4] |
| Lowest API price (per second, with audio) | from $0.05/s (Lite)[3] | No public data |
| Cheapest paid consumer plan | $19.99/mo Google AI Pro $99.99/mo Google AI Ultra[9] | No public data |
| Free tier | Gemini API free tier: Not available for Veo 3.1 Standard, Fast or Lite (paid tier only)[3] | No public data |
| Max length, single generation | 8 s[1] | 30 s[4] |
| Max resolution | 4k (8s length only)[1] | 480p / 720p (BytePlus ModelArk plans)[8] |
| Native audio | Yes[2] | Yes[4] |
| Image-to-video | Yes[2] | Yes[4] |
| Public API | Yes[7] | Yes (BytePlus ModelArk)[8] |
Pricing
Depends — not enough comparable public data| Metric | Veo 3.1 | Seedance 2.5 | Leader |
|---|---|---|---|
| API price, 720p with audio (per second) | $0.40/s Standard $0.10/s Fast $0.05/s Lite[3] | No public data | — |
| API price, 1080p with audio (per second) | $0.40/s Standard $0.12/s Fast $0.08/s Lite[3] | No public data | — |
| API price, 4K (per second) | $0.60/s Standard $0.30/s Fast[3] | No public data | — |
| API billing basis | Per second of successfully generated video, with audio included by default; price varies by tier (Standard / Fast / Lite) and by output resolution (720p / 1080p / 4k)[3] | No public data | — |
| Cheapest paid consumer plan | $19.99/mo Google AI Pro $99.99/mo Google AI Ultra[9] | No public data | — |
| Consumer plan ladder | Google AI Pro $19.99/month (1,000 Google Flow Credits, limited trial of Veo 3.1 Lite); Google AI Ultra from $99.99/month (10,000 or 25,000 Google Flow Credits, full Veo 3.1)[9] | No public data | — |
| In-app cost per second | No public data | Dreamina app: $0.035 per generated second at 720p (Basic plan, first-month promotional price); $0.047 per second (Super annual, first-year price)[10] | — |
| Free tier | Gemini API free tier: Not available for Veo 3.1 Standard, Fast or Lite (paid tier only)[3] | No public data | — |
Veo 3.1 leads 0 rows, Seedance 2.5 leads 0, 0 tied; rows marked — have no comparable data or no objective “better”.
Output specs
Winner: Seedance 2.5| Metric | Veo 3.1 | Seedance 2.5 | Leader |
|---|---|---|---|
| Max length, single generation | 8 s[1] | 30 s[4] | Seedance 2.5 |
| Max length with extensions | 148 s (with extensions, 720p)[1] | No public data | — |
| Max output resolution | 2160p (4K) (Standard)[1] | 720p (BytePlus ModelArk plans)[8] | Veo 3.1 |
| Frame rate | 24 fps[1] | No public data | — |
| Aspect ratios | 16:9 (default), 9:16[1] | No public data | — |
| Max reference inputs per generation | 3[1] | 50 (images + video + audio)[4] | Seedance 2.5 |
Veo 3.1 leads 1 row, Seedance 2.5 leads 2, 0 tied; rows marked — have no comparable data or no objective “better”.
Features
Tie| Metric | Veo 3.1 | Seedance 2.5 | Leader |
|---|---|---|---|
| Native audio | Yes[2] | Yes[4] | Tie |
| Lip-sync | No public data | No public data | — |
| Text-to-video | Yes[2] | Yes[4] | Tie |
| Image-to-video | Yes[2] | Yes[4] | Tie |
| First / last frame control | Yes[7] | No public data | — |
| Multi-shot in one generation | No public data | Yes[4] | — |
| Video editing | Yes (Lite)[11] | Yes (timestamp-level)[4] | Tie |
| Video extension | Yes[2] | Yes (multi-round)[4] | Tie |
Veo 3.1 leads 0 rows, Seedance 2.5 leads 0, 5 tied; rows marked — have no comparable data or no objective “better”.
Access
Tie| Metric | Veo 3.1 | Seedance 2.5 | Leader |
|---|---|---|---|
| Status | Paid preview in the Gemini API (Google's current model list names veo-3.1-generate-preview and veo-3.1-lite-generate-preview as Preview models; Veo 3.1 Fast rates are still published on the pricing page); also served via Vertex AI, the Gemini app and Flow[7] | Generally available — API fully available to clients on BytePlus ModelArk; live in Jimeng Web and Doubao Pro[8] | — |
| Public API | Yes[7] | Yes (BytePlus ModelArk)[8] | Tie |
| Where to use it | Gemini API in Google AI Studio and Vertex AI; Gemini app; Flow[7] | Jimeng and Doubao apps; Dreamina (international consumer app); API on BytePlus ModelArk[4] | — |
| Regional limits | In EU, UK, CH and MENA locations, allow_adult is the only permitted personGeneration value[1] | No public data | — |
| Release date | 2025-10-15[7] | 2026-07-31[4] | — |
Veo 3.1 leads 0 rows, Seedance 2.5 leads 0, 1 tied; rows marked — have no comparable data or no objective “better”.
Public leaderboards
Winner: Seedance 2.5| Metric | Veo 3.1 | Seedance 2.5 | Leader |
|---|---|---|---|
| LMArena text-to-video score | 1364 ±14 (veo-3.1-audio)[5] | 1474 ±9 (dreamina-seedance-2.5-720p)[5] | Seedance 2.5 |
| LMArena text-to-video rank (of 48) | #12 (veo-3.1-audio)[5] | #7 (dreamina-seedance-2.5-720p)[5] | — |
| LMArena image-to-video score | 1398 ±10 (veo-3.1-audio)[6] | 1477 ±7 (dreamina-seedance-2.5-720p)[6] | Seedance 2.5 |
| LMArena image-to-video rank (of 48) | #13 (veo-3.1-audio)[6] | #4 (dreamina-seedance-2.5-720p)[6] | — |
| VGI-Bench overall score | 32.0[12] | No public data | — |
Veo 3.1 leads 0 rows, Seedance 2.5 leads 2, 0 tied; rows marked — have no comparable data or no objective “better”.
How rows are judged: a lower price, a longer clip, a higher resolution or frame rate, more reference inputs and a supported feature win. Prices of a discontinued product are shown as historical and cannot win. LMArena scores are human-preference ratings (board snapshot dated 2026-09-21, read 2026-10-07); scores whose ± intervals overlap count as a tie. VGI-Bench is a research benchmark of visual reasoning. A row marked — has no comparable data.
Where each model leads
Where Veo 3.1 leads
Everything documented, everything priced. Veo 3.1 publishes its frame rate (24fps), resolutions, extension limits and first-and-last-frame control[1], plus a Fast tier at $0.10/s[3] per 720p second[3]. Consumer access runs through Google AI Pro at $19.99/mo[9] with only a limited Veo 3.1 Lite trial, or Google AI Ultra from $99.99/mo for the full model[9]. If that ladder is steep, AIContentDrop's credit plans include Veo 3.1 from $19/mo[13].
Where Seedance 2.5 leads
References, length and in-place edits. Up to 30 images, 10 video clips and 10 audio clips can steer one generation[4], output arrives as logically connected multi-shot sequences, and timestamp-level editing fixes a section without re-rendering the rest. You can try Seedance 2.5 as served on AIContentDrop ; for how it differs from its own stablemates, read our Seedance 2.5, Seedance 2.0 and Minimax H3 notes .
Which should you choose?
Pick Veo 3.1 if…
you deliver 4K, need a fixed frame rate, or bill per second: every Veo 3.1 price and output spec sits in public Google documentation[3].
Frequently asked questions
Which model wins on public leaderboards?
Seedance 2.5, on both boards that list the pair. LMArena text-to-video: rank #7[5] (1474[5]) against rank #12[5] (1364[5]) for veo-3.1-audio. LMArena image-to-video: rank #4[6] (1477[6]) against rank #13[6] (1398[6]).
Which produces higher-resolution video?
Veo 3.1 on the published record: 1080p and 4K for 8-second clips, 720p otherwise[1]. Seedance 2.5's official plan cards list only 480p and 720p[8]; there is no public data for anything higher.
Which is cheaper per second?
Not directly comparable. Google publishes Veo 3.1 rates from $0.05/s[3] (Lite, 720p) to $0.60/s[3] (Standard, 4K). ByteDance publishes no per-second API price for Seedance 2.5; the closest first-party figure is Dreamina's promotional $0.035 per generated second at 720p[10], a consumer app rate.
Evidence
Only official sources are cited: vendor pages, and leaderboard publishers for their own rankings. Anything we could not find on such a page is shown as “No public data”.
- Gemini API — Generate videos with Veo — Google. ai.google.dev/gemini-api/docs/veo. Checked 2026-10-06
- Gemini API — Video generation docs index — Google. ai.google.dev/gemini-api/docs/video. Checked 2026-10-06
- Gemini API — Pricing (Veo 3.1 Standard / Fast / Lite) — Google. ai.google.dev/gemini-api/docs/pricing. Checked 2026-10-06
- Introducing Seedance 2.5 — One-Take Creation, Flexible Referencing — ByteDance Seed. seed.bytedance.com/en/blog/one-take-creation-flexible-referencing-i…. Checked 2026-10-06
- Text-to-Video Arena leaderboard — LMArena (arena.ai). arena.ai/leaderboard/text-to-video. Checked 2026-10-06
- Image-to-Video Arena leaderboard — LMArena (arena.ai). arena.ai/leaderboard/image-to-video. Checked 2026-10-06
- Introducing Veo 3.1 and new creative capabilities in the Gemini API — Google Developers Blog. developers.googleblog.com/introducing-veo-3-1-and-new-creative-capa…. Checked 2026-10-06
- Dreamina-seedance-2.5 API Now Fully Available to Clients — BytePlus (ByteDance). www.byteplus.com/en/product/seedance. Checked 2026-10-06
- Google AI plans (Google One, US pricing) — Google. one.google.com/intl/en_us/about/google-ai-plans. Checked 2026-10-06
- Seedance 2.5 pricing (Dreamina) — Dreamina (ByteDance / CapCut). dreamina.capcut.com/seedance/seedance-2-5-pricing-2026. Checked 2026-10-06
- Gemini API — Models list (Veo 3.1, Veo 3.1 Lite) — Google. ai.google.dev/gemini-api/docs/models. Checked 2026-10-06
- VGI-Bench leaderboard (EMNLP 2026) — VGI-Bench research team. hexuan21.github.io/VGI-Bench. Checked 2026-10-06
- AIContentDrop — Pricing — AIContentDrop. aicontentdrop.com/pricing. Checked 2026-10-07