Updated November 8, 2026. The short answer, by job:
- Best quality right now: a split decision. Gemini Omni 1.2 Flash (Google, released November 5) leads the Arena leaderboard and costs about $0.07 per second at 720p. Wan 3.0 (Alibaba) leads Artificial Analysis's text-to-video-with-audio board at about $0.083 per second.
- Best value with sound: Gemini Omni 1.2 Flash ($0.07/s, up to 15 s), MiniMax H3 ($0.08/s at 768P) and Wan 3.0 ($0.083/s at 720P). All three include audio in the price.
- Animating a product photo: Seedance 2.0 ($0.15/s) and Kling 3.0 or the new Kling 4.0 ($0.094–0.157/s). All take a first frame and an optional last frame. Test them with your own photo.
- Longest single shot: Kling 4.0, Seedance 2.5 and Wan 3.0, all up to 30 seconds per generation.
- Gone or moving: Sora 2 was removed from OpenAI's API on September 24. Veo 3.1 left the Gemini API on October 22 and now lives on Vertex AI, Replicate and Runway; fal drops it on November 17.
Want to see what these models do with your product rather than a demo reel? Upload one product photo, describe the shot, and get a clip back in Teamday's video studio.
Make your first video from a product photo →Real Work First: A 30-Second Film Made in Teamday
This is a finished, assembled video, not a single best-case clip. A second example, Dawn Delivery (15 s), was built from generated source frames, image-to-video, music and assembly. You can read the director's note and production inputs or watch the 15-second cut. The whole workflow is described on the short-form video job page.
For downloadable checkpoints, hardware requirements and model-license differences, see our Hugging Face self-hosted video guide. If you want one API key for many of these models, our fal.ai vs Replicate comparison prices the same video models on both platforms.
November 2026 Comparison: Length, Audio, Price per Second
Prices are official API list prices from the model makers, checked on November 8, 2026, at 720p unless noted. "Audio" means the API can return sound generated with the video. Swipe the table sideways on mobile.
| Model | Maker | Max clip | Audio | Image-to-video | API $/sec (720p) | Best for | In Teamday |
|---|---|---|---|---|---|---|---|
Gemini Omni 1.2 Flash (gemini-omni-1.2-flash) | 15 s (extend to 30 s) | Yes | Yes | ~$0.07 | Top quality per dollar, conversational editing | No | |
| Wan 3.0 | Alibaba | 30 s | Yes | Yes (first / last frame) | $0.083 (US) · $0.10 (Singapore) | Long shots with sound on a budget | No (Wan 2.7 is) |
| MiniMax H3 | MiniMax | 15 s | Yes, stereo | Yes | $0.08 (768P) · $0.13 (2K) | Cheap high resolution, open weights | No |
| Seedance 2.0 / 2.0 Fast / 2.0 Mini | ByteDance | 15 s | Yes | Yes (first / last frame) | ~$0.15 / $0.12 / $0.076 | Product and people image-to-video | Yes (2.0, 2.0 Fast) |
| Seedance 2.5 | ByteDance | 30 s | Yes | Yes (up to 30 image references) | ~$0.23 | Long, reference-heavy scenes | Yes |
| Kling 4.0 / 4.0 Flash | Kuaishou | 30 s / 20 s | Yes, stereo | Yes | ~$0.157 / ~$0.094 | 30-second shots from a still | No |
| Kling 3.0 / 3.0 Omni | Kuaishou | 15 s | Optional | Yes (start / end frame) | ~$0.094 silent · ~$0.14 audio | Cinematic motion from a still | Yes (Kling 3 Pro) |
| Grok Imagine Video 1.5 / 1.5 Lite | xAI | 15 s | Yes | Yes | $0.14 / $0.03 (+$0.01 per input image) | Fast short-form clips | Yes (1.5, 720p) |
| HappyHorse 1.1 | Alibaba | Not stated (1.0: 15 s) | Not stated | Yes | $0.124 (US) · $0.14 (Singapore) | Image-to-video in the Alibaba stack | 1.0 only |
| FLUX 3 Video | Black Forest Labs | 20 s | Yes, included | Yes | $0.15 (HD) · $0.24 (FHD) · $0.06 (Draft) | Long clips with audio from an image lab | No |
| Veo 3.1 Standard / Fast / Lite | Google (Vertex AI only) | 8 s (extendable) | Yes, always on | Yes | $0.40 / $0.10 / $0.05 | 4K output (Standard, Fast) | No |
| LTX-2.5 Fast / Pro | Lightricks | 20 s / 10 s | Yes | Yes | $0.09 / $0.12 | Open weights, up to 4K | No |
| Runway Gen-4.5 | Runway | 10 s | No (API) | Yes | $0.12 | Hands-on creative editing | No |
| Luma Ray 3.2 | Luma | 10 s (API) | Not documented | Yes (keyframes) | $0.06 · $0.24 at 1080p | HDR and EXR delivery for post-production | No |
| Pika 2.5 | Pika | Not stated | Not stated | Yes | $0.04 · $0.06 at 1080p | Cheapest listed API rate | No |
| Sora 2 | OpenAI | — | — | — | Removed Sept 24, 2026 | Nothing new | No |
Not in the table: Midjourney Video (no public API price), Adobe Firefly Video (no public per-second API price; the Firefly app hosts partner models such as Veo, Kling, Runway and Luma), Tencent HunyuanVideo 1.5 (open weights, 720p native, no official hosted API price), Utopai X and Meta Muse Video (both appear on leaderboards, but we could not find public API pricing, so we do not rank them here).
What the Blind-Vote Leaderboards Say
Two public arenas rank video models by head-to-head human votes. They no longer agree on the winner (checked November 8, 2026):
| Rank | Artificial Analysis, text-to-video with audio (v2.1 board) | Arena, text-to-video (dated Nov 6) |
|---|---|---|
| 1 | Wan 3.0 (1172) | Gemini Omni 1.2 Flash (1541, preliminary) |
| 2 | Utopai X (1167) | Gemini Omni 1.1 Flash (1514) |
| 3 | MiniMax H3, 768p (1160) | Gemini Omni Flash preview (1509) |
| 4 | FLUX 3 Max (1156) | FLUX 3 Video (1491, preliminary) |
| 5 | Seedance 2.5 (1155) | Grok Imagine Video 1.5 (1486, preliminary) |
Seedance 2.0 sits 6th on Arena and 10th on Artificial Analysis. Both boards put Kling 3.0 and Veo 3.1 well outside the top ten: Veo 3.1 is 23rd on Artificial Analysis and 14th on Arena, and Veo 3.1 Standard still costs nearly six times as much as Omni 1.2 Flash at 720p. Kling 4.0 was not yet scored on either board when we checked. Artificial Analysis rescaled its scores since September (its v2.1 board), so do not compare these numbers with older ones.
Leaderboards measure how good a single clip looks from a text prompt. They do not measure whether your product label survives, whether shot 3 matches shot 2, or what the finished ad costs. That is what the rest of this guide covers.
The Models, One by One
1. Gemini Omni 1.2 Flash: 15 seconds, 30% cheaper
Google released Gemini Omni 1.2 Flash as gemini-omni-1.2-flash on November 5, 2026. It generates 3–15 second clips (up from 10 seconds in Omni 1.1) at 720p by default, with 1080p and 4K upscaled output at 24 fps, and a synchronized audio track with improved lip sync. You can extend a clip in 3–10 second turns up to 30 seconds total, and edit a clip in a multi-turn conversation. Pricing is per token: $12 per million video output tokens at 5,792 tokens per second of 720p video, which Google rounds to about $0.07 per second, or $1.04 for a 15-second clip. 1080p costs about $0.104 per second and 4K about $0.21.
Two housekeeping notes. The previous gemini-omni-1.1-flash ($0.10/s) routes to 1.2 and is billed at 1.2 rates from December 3, 2026. And Google's Gemini API docs moved from ai.google.dev to aistudio.google.com, so old bookmarks redirect.
Teamday: not integrated yet.
2. Wan 3.0: the audio leaderboard's number one
Alibaba's Wan 3.0 (wan3.0-video) produces 2–30 second clips at up to 1080P with dialogue, music and sound effects on by default, from text, a first frame, first and last frames, or references. Audio does not change the price. It costs $0.0825 per second at 720P in the US region and $0.10 in Singapore; 1080P is $0.165 and $0.20. The faster Wan 3.0 Prime costs $0.127 (US) or $0.14 (Singapore) at 720P. Alibaba's Model Studio pages say nothing about open weights for 3.0, unlike Wan 2.x. Nothing newer than Wan 3.0 had appeared in the Model Studio model list, last updated November 2, when we checked.
Teamday: runs the previous generation, Wan 2.7, for image-to-video at 720p. Wan 3.0 is not integrated yet.
3. MiniMax H3: 2K with stereo audio for $0.13 a second
MiniMax H3 generates 4–15 second clips with native stereo audio and supports image references. Official pricing is $0.08 per second at 768P and $0.13 at 2K, still the cheapest 2K-with-audio rate we found. The pricing page now also lists H3-Max ($0.08 at 768P, $0.05 at 480P), H3-Highspeed ($0.06 at 768P) and H3-Lite ($0.04 at 768P), though MiniMax's model documentation describes only H3 itself, so check duration and audio support for the variants before you budget on them. The 768p base weights are on Hugging Face under a community license; the 2K regeneration model is not open. Hailuo 2.3 and 02 are now labelled legacy.
Teamday: not integrated.
4. Seedance 2.0, 2.0 Mini and 2.5: the image-to-video workhorses
ByteDance's Seedance line is in the top ten on both leaderboards. Seedance 2.5 runs a single generation to 30 seconds and accepts up to 30 image references, so a product, a person and a location can stay consistent. Seedance 2.0 runs to 15 seconds and up to 4K. All generate audio with the video.
BytePlus ModelArk pricing (last updated October 23, 2026) is per token: width × height × fps × seconds / 1024. Audio does not change the price. For a 16:9 clip at 720p with no video input, BytePlus's own examples work out to:
- Seedance 2.0: $7.00 per million tokens, about $0.15 per second at 720p ($0.07 at 480p, $0.37 at 1080p).
- Seedance 2.0 Fast: $5.60 per million, about $0.12 per second at 720p (480p and 720p only).
- Seedance 2.0 Mini (new on the price list): $3.50 per million, about $0.076 per second at 720p.
- Seedance 2.5: $10.70 per million, $0.231 per second at 720p and $0.57 at 1080p.
Seedance 1.5 Pro retires on November 11, 2026. We have seen no official Seedance 3.0; sites claiming one are not ByteDance's. We covered the original launch in our Seedance 2 analysis.
Teamday: Seedance 2.0 and 2.0 Fast (4–15 s) and Seedance 2.5 (4–30 s, when your ModelArk account has access) run at 480p, 720p or 1080p. They accept a first frame and an optional last frame. Native audio is off by default so that voiceover and music can be mixed in assembly. Agents can turn it on for a clip. Seedance 2.0 Mini is not integrated.
5. Kling 4.0 and 3.0: 30-second shots arrive
Kuaishou launched Kling 4.0 on October 15, 2026. It generates 3–30 second clips with two-channel stereo audio at 720p or 1080p, with 4K and 10-bit HDR listed as coming soon. Kling 4.0 Flash is the cheaper 720p-only version with 3–20 second clips. Kling 3.0 and 3.0 Omni (references and editing) remain on the API with 3–15 second clips, start and end frames, and audio switchable between native, original and off.
Kling's developer pricing is in units per second, and Kling does not print a per-unit dollar price; its $510 pack buys 3,250 units, so one unit is about $0.157 at list. That gives:
- Kling 4.0: 1.0 unit per second at 720p (about $0.157/s) and 1.5 units at 1080p (about $0.235/s), audio included.
- Kling 4.0 Flash: 0.6 units per second at 720p (about $0.094/s).
- Kling 3.0: 0.6 units per second silent or 0.9 with native audio at 720p (about $0.094 and $0.14/s), 0.8 or 1.2 at 1080p, and 3.0 units at 4K.
fal, Replicate and Runway all resell Kling 4.0 at $0.157 per second at 720p, which matches the pack math. Test how well it keeps your product, logo area and first frame, not just how dramatic the motion looks.
Teamday: Kling 3 Pro text-to-video and image-to-video through FAL, 3–15 second clips, with audio optional. Kling 2.6 Pro remains available as a cheaper fallback. Kling 4.0 is not integrated yet.
6. Grok Imagine Video 1.5 and 1.5 Lite: fast short-form
xAI's Grok Imagine Video 1.5 generates 1–15 second clips with an audio track by default and lip sync, from text, an image or up to 14 reference images. xAI's pricing page now lists per-resolution rates: $0.08 per second at 480p, $0.14 at 720p and $0.25 at 1080p, plus $0.01 per input image. The new Grok Imagine Video 1.5 Lite costs $0.02 at 480p, $0.03 at 720p and $0.14 at 1080p (upscaled from 720p). xAI lists no video model newer than 1.5.
Teamday: grok-imagine-video-1.5 for image-to-video and grok-imagine-video for text-to-video, 1–15 seconds at 720p, with the provider's audio. The Lite tier is not integrated.
7. HappyHorse 1.1: Alibaba's image-to-video tier
HappyHorse 1.1 generates clips at up to 1080P from text, an image or references. It costs $0.124 per second at 720P in the US region and $0.14 in Singapore ($0.062 and $0.07 at 480P). Alibaba's 1.1 pages do not state a clip length or whether audio is returned; the 1.0 page says 3–15 seconds.
Teamday: HappyHorse 1.0 image-to-video at 720p, 2–15 seconds. If you start from text only, Teamday generates the first frame for you, which adds an image cost.
8. FLUX 3 Video: Black Forest Labs enters video
Black Forest Labs, the FLUX image lab, now sells FLUX 3 Video with clips up to 20 seconds and audio included at no extra charge. Pricing is by output class: $0.15 per second at HD (0.5–1 megapixel, for example 1280×704), $0.24 at FHD, $0.41 at 4K, and $0.06 for Draft. FLUX 3 Video is 4th on Arena and FLUX 3 Max is 4th on Artificial Analysis's audio board. fal and Replicate both resell it at $0.85 per 5-second 720p clip ($0.17/s).
Teamday: not integrated.
9. Veo 3.1: Vertex AI only, no longer the value pick
Google shut down the Veo 3.1 Preview models in the Gemini API on October 22, 2026 and points developers to Gemini Omni 1.2 Flash. The GA Veo 3.1 models remain on Vertex AI (Gemini Enterprise Agent Platform), and Replicate and Runway resell them. Veo 3.1 generates 4, 6 or 8 second clips (extendable on Standard and Fast) at up to 4K with audio always on, first and last frames, and up to three reference images. Vertex AI list prices, which we checked on November 7 and which Replicate matches:
- Standard: $0.40 per second at 720p or 1080p with audio, $0.60 at 4K; $0.20 silent.
- Fast: $0.10 at 720p with audio, $0.12 at 1080p.
- Lite: $0.05 at 720p with audio, $0.08 at 1080p; $0.03 silent. No extension, no 4K.
Veo 3.1 Lite is still the cheapest audio-video rate in this guide, but only if you are on Google Cloud or Replicate. fal removes its Veo 3.1 endpoints on November 17, 2026. Veo 3.1 Standard is hard to justify now that Omni 1.2 Flash ranks higher at under a fifth of the price.
Teamday: not integrated.
10. LTX-2.5: open weights with audio
Lightricks' LTX-2.5 generates video and audio together at up to 4K. The API costs $0.09 per second at 720p on Fast (6–20 second clips at 720p and 1080p) and $0.12 on Pro (6, 8 or 10 seconds; $0.17 at 1080p, $0.39 at 4K). The 22B weights are on Hugging Face under a community license that is free below $10M in annual revenue, which matters if you need to run video generation on your own infrastructure.
Teamday: not integrated.
11. Runway Gen-4.5, Luma Ray 3.2 and Pika 2.5: creative tools, silent or undocumented audio
Runway Gen-4.5 costs 12 credits per second at $0.01 per credit, so $0.12 per second, for 2–10 second clips at 1280×720 or 720×1280. The API has no audio parameter for Gen-4.5; its audio toggle exists only for the models Runway resells (Veo 3.1, Kling, Seedance, Wan). Runway has shipped no new model of its own since Gen-4.5 (December 2025), but its API now resells Kling 4.0, Seedance 2.5, Wan 3.0, Gemini Omni 1.2 Flash and Veo 3.1 next to it.
Luma Ray 3.2 is priced per 5-second clip: $0.30 at 720p and $1.20 at 1080p, or $0.06 and $0.24 per second, with HDR at double. The API caps clips at 10 seconds and supports start and end frames plus up to 64 guide images. Luma's docs do not mention audio output, and the pricing page warns that video rates may change before general availability.
Pika 2.5 now publishes an API catalog price of $0.04 per second at 720p and $0.06 at 1080p, the lowest listed rate in this guide. The catalog does not state clip length or audio, so treat it as a draft tool until you have tested it.
Teamday: none of the three is integrated.
What happened to OpenAI Sora?
OpenAI removed sora-2, sora-2-pro and the Videos API on September 24, 2026. The deprecation page lists no replacement, and OpenAI had announced no successor video model when we checked. Any "best video model" list that still recommends Sora is out of date.
Worked Example: What a 30-Second Ad Costs
A 30-second ad is typically four to six shots. The model cost is the price per second times the seconds you generate, not the seconds you keep. Below is list price for 30 seconds of 720p footage. The 3× column assumes you generate three takes for every second you use; that is an illustrative assumption, not a measured rate, so replace it with your own rejection rate.
| Model (720p) | $/sec | 30 s, one take each | 30 s, 3 takes each |
|---|---|---|---|
| Grok Imagine Video 1.5 Lite (audio) | $0.03 | $0.90 + $0.01 per input image | $2.70 + image fees |
| Veo 3.1 Lite, Vertex AI (audio) | $0.05 | $1.50 | $4.50 |
| Luma Ray 3.2 (audio not documented) | $0.06 | $1.80 | $5.40 |
| Gemini Omni 1.2 Flash (audio) | ~$0.07 | $2.09 | $6.26 |
| Seedance 2.0 Mini, 16:9 (audio) | ~$0.076 | $2.27 | $6.80 |
| MiniMax H3, 768P (audio) | $0.08 | $2.40 | $7.20 |
| Wan 3.0, US region (audio) | $0.0825 | $2.48 | $7.43 |
| LTX-2.5 Fast (audio) | $0.09 | $2.70 | $8.10 |
| Kling 3.0 (silent) | ~$0.094 | $2.83 | $8.48 |
| Runway Gen-4.5 (silent) | $0.12 | $3.60 | $10.80 |
| Kling 3.0 (audio) | ~$0.14 | $4.24 | $12.72 |
| Grok Imagine Video 1.5 (audio) | $0.14 | $4.20 + $0.01 per input image | $12.60 + image fees |
| FLUX 3 Video, HD (audio) | $0.15 | $4.50 | $13.50 |
| Seedance 2.0, 16:9 (audio) | ~$0.15 | $4.54 | $13.61 |
| Kling 4.0 (audio) | ~$0.157 | $4.71 | $14.13 |
| Seedance 2.5, 16:9 (audio) | ~$0.23 | $6.93 | $20.80 |
| Veo 3.1 Standard, Vertex AI (audio) | $0.40 | $12.00 | $36.00 |
Three conclusions:
- Generation is now the cheap part. Even with three takes per shot, most models cost less than $15 for a 30-second ad. Your time writing the brief, reviewing takes and fixing continuity costs more.
- Silent models are not cheaper if you then buy audio. Runway leaves you to add sound separately. Omni 1.2 Flash, Wan 3.0 and MiniMax H3 include it at a similar or lower price.
- Rejection rate matters more than list price. A $0.15/s model that keeps your product intact on the first take beats a $0.07/s model that needs five attempts.
How to Choose
Answer three questions in order:
- Do you start from a photo of a real product or person? If so, image-to-video fidelity comes first. Test Seedance 2.0, Kling 3.0 or 4.0, and Wan 3.0 with the same photo, and reject any take where the label, logo or face changes.
- Do you need sound in the clip? Omni 1.2 Flash, Wan 3.0, MiniMax H3, Seedance, Kling, Grok, FLUX 3 Video, LTX-2.5 and Veo generate it. If you will add voiceover and music in editing anyway, generate silent clips on Kling 3.0 at 720p for about $0.094/s.
- How long is one shot? For shots over 15 seconds without cuts, Kling 4.0, Seedance 2.5 and Wan 3.0 (30 s), FLUX 3 Video and LTX-2.5 Fast (20 s), and Kling 4.0 Flash (20 s) do it in one generation. Otherwise, extend Omni 1.2 Flash (to 30 s) or Veo.
Then run a fair test. Split one 20–30 second concept into the same shot list, give every model identical prompts and source frames, and score each take:
| Criterion | What to measure |
|---|---|
| First-frame fidelity | Does the clip keep the approved composition, product and face? |
| Motion | Are bodies, objects, cloth, liquid and camera movement physically believable? |
| Continuity | Do identity, wardrobe, product, location and lighting survive across shots? |
| Prompt following | Did the requested action and camera move happen? |
| Audio | Are dialogue, ambience, effects and lip sync usable? |
| Rejection rate | Rejected takes divided by all takes |
| Cost per approved video | Model cost plus review and editing time, divided by approved deliverables |
Do not compare one model's best take with another's first take. Record every attempt.
Doing This in Teamday
Teamday's video studio and Reel, the AI video producer, handle the full job: brief, storyboard, source frames, clips, voice, music and final cut. You pick the model per shot from the routes Teamday supports today:
| Teamday route | Input | Clip length | Audio in Teamday |
|---|---|---|---|
| Seedance 2.0 / 2.0 Fast | Text, optional first and last frame | 4–15 s, 480p–1080p | Off by default; can be requested |
| Seedance 2.5 | Text, optional first and last frame | 4–30 s, 480p–1080p | Off by default; can be requested (needs ModelArk access) |
| Kling 3 Pro (FAL) | Text or image | 3–15 s | Optional |
| Kling 2.6 Pro, Wan 2.6 (FAL) | Text (Kling also image) | Short clips | Added in assembly |
| Grok Imagine Video / 1.5 (xAI) | Text / image | 1–15 s, 720p | Provider audio |
| HappyHorse 1.0, Wan 2.7 (Alibaba) | Image (auto-generated from text if needed) | 2–15 s, 720p | Added in assembly |
Every generated clip records the exact model ID it was made with, so when a better model ships you know which shots to regenerate.
Run the fair test on your own product instead of reading about it. Upload a product photo, and Reel storyboards the shots, generates takes on Seedance or Kling, and assembles the cut with voice and music for you to approve.
Animate your product photo in the video studio →Frequently Asked Questions
What is the best AI video generator in November 2026?
The two blind-vote leaderboards disagree. Arena puts Google's Gemini Omni 1.2 Flash first; Artificial Analysis's text-to-video-with-audio board puts Alibaba Wan 3.0 first, with MiniMax H3, FLUX 3 Max and Seedance 2.5 close behind. For animating a product photo, Seedance 2.0 and Kling 3.0 or 4.0 remain the models to test. The right pick depends on clip length, audio, resolution and price per second.
What is the cheapest good AI video model?
At official API list prices checked on November 8, 2026:
- Veo 3.1 Lite: $0.05 per second at 720p, with audio, on Vertex AI or Replicate.
- Gemini Omni 1.2 Flash: about $0.07 per second at 720p.
- Seedance 2.0 Mini: about $0.076 per second at 720p.
- MiniMax H3: $0.08 per second at 768P.
- Wan 3.0: about $0.083 per second at 720P (US region).
Thirty seconds of footage costs roughly $1.50 to $2.50 on these models, before retakes.
How much does a 30-second AI video ad cost to generate?
At 720p list prices, 30 seconds of generated footage costs about $1.50 (Veo 3.1 Lite) to $6.93 (Seedance 2.5), and $12 on Veo 3.1 Standard. If you generate three takes for every second you keep, multiply by three: roughly $4.50 to $21, or $36 on Veo 3.1 Standard. Voice, music, editing and review time come on top.
Which AI video models generate native audio?
Gemini Omni 1.2 Flash, Veo 3.1, Wan 3.0, MiniMax H3, Seedance 2.0 and 2.5, Kling 3.0 and 4.0, Grok Imagine Video 1.5, LTX-2.5 and FLUX 3 Video generate audio with the video. Runway Gen-4.5's API has no audio option, and Luma's Ray 3.2 docs do not mention audio.
Is OpenAI Sora still available?
No. OpenAI removed sora-2, sora-2-pro and the Videos API on September 24, 2026, and its deprecation page lists no replacement.
Can I still use Google Veo 3.1 through an API?
Only on Vertex AI (Gemini Enterprise Agent Platform), Replicate and Runway. The Veo 3.1 Preview models in the Gemini API shut down on October 22, 2026, and Google points developers to Gemini Omni 1.2 Flash instead. fal removes its Veo 3.1 endpoints on November 17, 2026.
Which AI video models can I use in Teamday?
Teamday's video studio runs Seedance 2.0, Seedance 2.0 Fast and Seedance 2.5 (ByteDance ModelArk), Kling 3 Pro and Kling 2.6 Pro (via FAL), Grok Imagine Video and Grok Imagine Video 1.5 (xAI), HappyHorse 1.0 and Wan 2.7 image-to-video (Alibaba), and Wan 2.6 (via FAL). Gemini Omni, Kling 4.0, Veo, Wan 3.0, MiniMax H3, FLUX 3 Video, Runway, Luma and LTX are not integrated yet.
Can I turn a product photo into a video?
Yes. All major models support image-to-video. In Teamday's video studio you upload a product photo, describe the motion, and it is used as the first frame of the clip.
Pick one product and one 30-second idea. Teamday's video studio takes you from photo to finished cut, and every clip records which model made it.
Start your first video →Related Guides
- Best AI image models 2026
- Best AI voice models 2026
- Best AI avatar models 2026
- fal.ai vs Replicate: the same video models priced on both
- AI image and video API providers compared
- AI video ads use case
- Teamday model catalog
Research date: November 8, 2026. Prices are official vendor API list prices from the pages linked above, in USD, without enterprise or promotional discounts; Veo 3.1 prices are Vertex AI list prices checked November 7, 2026. Kling dollar figures are derived from its unit-pack pricing. Artificial Analysis showed no board-wide update date; Arena was dated November 6. Model availability, regions and prices change often, so confirm the model ID, region, duration and audio setting before you commit a production budget.
