
11 Best Text-to-Video APIs in 2026, Tested
Summary
Seedance 2.5 leads the leaderboard of the best text-to-video models available.
Wan 3.0 is second, with a much lower price.
Gemini Omni 1.2 Flash, MiniMax H3 and MiniMax H3 Max tie for third, each for a fraction of Seedance 2.5’s price per second.
Seedance 2.0, Gemini Omni 1.1 Flash, Kling 4.0, Grok Imagine Video 1.5, HappyHorse 1.0 and Veo 3.1 follow, in that order.
All of these models are available on Runway Dev.
A text-to-video model has nothing to lean on but the prompt: no first frame fixes the look, no reference fixes the product or the cast. So the best text-to-video model is the one that turns a prompt alone into the clip it describes.
What are the best text-to-video APIs in 2026?
Seedance 2.5 from ByteDance leads my text-to-video test, Wan 3.0 from Alibaba is a tenth of a point behind, and three models share third place. Every model here runs on Runway Dev, and the last column is its price for one second of video there.
| Rank | Model | Lab | Best for | Score | Price on Runway Dev |
|---|---|---|---|---|---|
| 1 | Seedance 2.5 | ByteDance | Long multi-shot scenes that follow every beat of the prompt | 4.3 | 68 credits ($0.68) per second at 1080p |
| 2 | Wan 3.0 | Alibaba | Long single takes with sound, close to the top for far less | 4.2 | 20 credits ($0.20) per second at 1080p |
| 3= | Gemini Omni 1.2 Flash | Generating, editing and extending in one model | 4.1 | 10 credits ($0.10) per second at 1080p | |
| 3= | MiniMax H3 | MiniMax | Cinematic scenes at 2K | 4.1 | 15 credits ($0.15) per second at 2K |
| 3= | MiniMax H3 Max | MiniMax | Cinematic results at the lowest price in the test | 4.1 | 8 credits ($0.08) per second at 768p |
| 6 | Seedance 2.0 | ByteDance | Multi-shot scenes for less than Seedance 2.5 | 3.8 | 40 credits ($0.40) per second at 1080p |
| 7 | Gemini Omni 1.1 Flash | Grounded, cinematic shots with native audio | 3.6 | 15 credits ($0.15) per second at 1080p | |
| 8 | Kling 4.0 | Kling | Clips up to 30 seconds with audio | 3.5 | 23.5 credits ($0.235) per second at 1080p |
| 9 | Grok Imagine Video 1.5 | xAI | Fast generation with native audio | 3.3 | 29 credits ($0.29) per second at 1080p |
| 10 | HappyHorse 1.0 | Alibaba | Native audio in ten output shapes | 2.9 | 30 credits ($0.30) per second at 1080p |
| 11 | Veo 3.1 | Short shots with audio, steered by a negative prompt | 2.4 | 40 credits ($0.40) per second (With audio) |
What makes a text-to-video model the best?
The best text-to-video model turns a prompt into exactly the clip it describes: every action in order, the same subject from start to end, motion with weight, the camera move the prompt asks for and sound that lands on the picture. Ten criteria decide it:
- Prompt adherence: every beat of the prompt happens, in order, with nothing dropped or added.
- Product fidelity: a product keeps the shape, color and lettering the prompt describes.
- Character consistency: the same face, hair and clothes from the first frame to the last.
- Performance: expressions, timing, eye-lines and gestures that feel directed.
- Movement and physics: bodies, fabric, liquids and objects move with weight and cause.
- Hands and handling: grips and gestures without fused or extra fingers.
- Camera: the requested move happens smoothly, with no jumps or cuts.
- Voice and lip sync: the exact lines, in sync, in distinct voices where there are two.
- Sound effects: effects and ambience timed to the picture, with music only where the prompt asks for it.
- Visual quality: detail and light, with no flicker or warping.
Every model ran the same ten briefs as the best AI video models, written as prompts with no image, and I scored each take from 1 to 5 on these criteria. The rank follows the average.
Prompt adherence matters most in text-to-video, because no first frame or reference image fixes the scene. Each model section below plays the same prompt, rendered at 8 seconds: one robot, one camera move and one jump over a rock.
Arri Alexa 65mm camera style, film grain slightly visible, shallow depth of field, cool teal shadows and warm highlights, slight camera shake. One continuous side-profile tracking shot, no cuts. An unbranded gray quadruped robot walks along a dirt trail through a tall pine forest at golden hour, framed on the left side of the frame, the camera moving alongside it at the same speed. A mossy rock comes into frame from the right. The robot slows, crouches and jumps over the rock, lands on all four legs and keeps walking at the same steady pace. The background is the forest, with sunlight through the trees. Audio: no music, only sound effects: the robot's motors whirring, its feet on the dirt, the thud of the landing, birds in the forest.
1. Seedance 2.5 follows a prompt beat by beat
Seedance 2.5 scored 4.3, the highest of the 11 text-to-video models in my test.
Price on Runway Dev: 68 credits ($0.68) per second at 1080p.
- Best for: long multi-shot scenes that follow every beat of the prompt.
- Longest clip: 30 seconds in one generation.
- Prompt length: up to 15,000 characters.
- Sound: generated with the picture.
- An 8-second clip: 544 credits ($5.44) on Runway Dev.
Call it with model: "seedance2_5" on image-to-video, text-to-video and video-to-video. The Seedance 2.5 API guide has the request in curl, Node.js and Python, and you can try it in the playground.
2. Wan 3.0 comes within a tenth of a point for far less
Wan 3.0 scored 4.2, one tenth behind Seedance 2.5, at a lower price per second.
Price on Runway Dev: 20 credits ($0.20) per second at 1080p.
- Best for: long single takes with sound.
- Longest clip: 30 seconds in one generation.
- Prompt length: up to 20,000 characters.
- Sound: generated in the same pass.
- An 8-second clip: 160 credits ($1.60) on Runway Dev.
Call it with model: "wan3" on image-to-video and text-to-video. The Wan 3.0 API guide has the request in curl, Node.js and Python, and you can try it in the playground.
3. Gemini Omni 1.2 Flash generates, edits and extends in one model
Gemini Omni 1.2 Flash scored 4.1, tied for third, and the same model also edits and extends a clip you already have.
Price on Runway Dev: 10 credits ($0.10) per second at 1080p.
- Best for: generating, editing and extending in one model.
- Longest clip: 15 seconds in one generation.
- Prompt length: up to 4,000 characters.
- Sound: native audio in the same pass.
- An 8-second clip: 80 credits ($0.80) on Runway Dev.
Call it with model: "gemini_omni_flash_1.2" on text-to-video, image-to-video and video-to-video, or try it in the playground and copy the request from there.
3. MiniMax H3 renders cinematic scenes at 2K
MiniMax H3 scored 4.1, tied for third, and renders at 2K as well as 768p.
Price on Runway Dev: 15 credits ($0.15) per second at 2K.
- Best for: cinematic scenes at 2K.
- Longest clip: 15 seconds in one generation.
- Prompt length: up to 6,000 characters.
- An 8-second clip: 120 credits ($1.20) on Runway Dev.
Call it with model: "hailuo3" on image-to-video, text-to-video and video-to-video, or try it in the playground and copy the request from there.
3. MiniMax H3 Max ties for third at the lowest price in the test
MiniMax H3 Max scored 4.1, tied for third, and costs the least per second of the 11 models.
Price on Runway Dev: 8 credits ($0.08) per second at 768p.
- Best for: cinematic results at the lowest price in the test.
- Longest clip: 15 seconds in one generation.
- Prompt length: up to 6,000 characters.
- Sound: always on, generated with the picture.
- An 8-second clip: 64 credits ($0.64) on Runway Dev.
Call it with model: "h3_max" on image-to-video and text-to-video, or try it in the playground and copy the request from there.
6. Seedance 2.0 makes multi-shot scenes for less than Seedance 2.5
Seedance 2.0 scored 3.8 and costs less per second than Seedance 2.5.
Price on Runway Dev: 40 credits ($0.40) per second at 1080p.
- Best for: multi-shot scenes for less than Seedance 2.5.
- Longest clip: 15 seconds in one generation.
- Prompt length: up to 10,000 characters.
- Sound: generated with the picture.
- An 8-second clip: 320 credits ($3.20) on Runway Dev.
Call it with model: "seedance2" on image-to-video, text-to-video and video-to-video. The Seedance 2.0 API guide has the request in curl, Node.js and Python, and you can try it in the playground.
7. Gemini Omni 1.1 Flash makes grounded shots with native audio
Gemini Omni 1.1 Flash scored 3.6 and generates sound in the same pass as the picture.
Price on Runway Dev: 15 credits ($0.15) per second at 1080p.
- Best for: grounded, cinematic shots with native audio.
- Longest clip: 10 seconds in one generation.
- Prompt length: up to 4,000 characters.
- An 8-second clip: 120 credits ($1.20) on Runway Dev.
Call it with model: "gemini_omni_flash_1.1" on text-to-video, image-to-video and video-to-video. The Gemini Omni 1.1 Flash API guide has the request in curl, Node.js and Python, and you can try it in the playground.
8. Kling 4.0 returns clips up to 30 seconds with audio
Kling 4.0 scored 3.5 and returns one of the longest clips in the test from one generation.
Price on Runway Dev: 23.5 credits ($0.235) per second at 1080p.
- Best for: long clips with audio.
- Longest clip: 30 seconds in one generation.
- Prompt length: up to 26,000 characters.
- Sound: included in every clip.
- An 8-second clip: 188 credits ($1.88) on Runway Dev.
Call it with model: "kling_4" on image-to-video, text-to-video and video-to-video. The Kling 4.0 API guide has the request in curl, Node.js and Python, and you can try it in the playground.
9. Grok Imagine Video 1.5 generates fast, with native audio
Grok Imagine Video 1.5 scored 3.3 and renders from 480p to 1080p with sound.
Price on Runway Dev: 29 credits ($0.29) per second at 1080p.
- Best for: fast generation with native audio.
- Longest clip: 15 seconds in one generation.
- Prompt length: up to 2,500 characters.
- An 8-second clip: 232 credits ($2.32) on Runway Dev.
Call it with model: "grok_imagine_1_5" on image-to-video and text-to-video, or try it in the playground and copy the request from there.
10. HappyHorse 1.0 makes clips with audio in ten output shapes
HappyHorse 1.0 scored 2.9 and offers ten aspect ratios, from vertical to widescreen.
Price on Runway Dev: 30 credits ($0.30) per second at 1080p.
- Best for: native audio in many output shapes.
- Longest clip: 15 seconds in one generation.
- Prompt length: up to 2,500 characters.
- An 8-second clip: 240 credits ($2.40) on Runway Dev.
Call it with model: "happyhorse_1_0" on image-to-video and text-to-video, or try it in the playground and copy the request from there.
11. Veo 3.1 makes short shots with audio, steered by a negative prompt
Veo 3.1 scored 2.4 and returns shots of 4, 6 or 8 seconds, with a negative prompt to keep things out of the frame.
Price on Runway Dev: 40 credits ($0.40) per second (With audio).
- Best for: short shots with audio, steered by a negative prompt.
- Longest clip: 8 seconds in one generation.
- Prompt length: up to 1,000 characters.
- An 8-second clip: 320 credits ($3.20) on Runway Dev.
Call it with model: "veo3.1" on image-to-video and text-to-video. The Veo 3.1 API guide has the request in curl, Node.js and Python, and you can try it in the playground.
How do you call every top text-to-video model with one API key?
Runway Dev serves all 11 models in this test behind one API key, one credit balance and one bill, so switching from Wan 3.0 to Seedance 2.5 means changing the model field of the request, with the same SDK and the same task flow. Create an account at dev.runway.com, make an API key and add credits.
import RunwayML from '@runwayml/sdk';
const client = new RunwayML(); // reads RUNWAYML_API_SECRET
const task = await client.textToVideo
.create({
model: 'seedance2_5', // or 'wan3', 'gemini_omni_flash_1.2'...
promptText:
'A quadruped robot jumps over a rock on a forest trail, the camera tracking alongside.',
ratio: '1920:1080',
duration: 8,
})
.waitForTaskOutput();
console.log(task.output[0]); // the video URL
A coding agent can write the integration for you. The Runway Dev plugin installs the Runway Dev MCP and the Runway Dev skills in Claude Code, Codex, Cursor and Grok Bot, so the agent sees the models your project can call, with their current prices and limits.
Which text-to-video model should you use for each request?
The right text-to-video model depends on the request: a long scene with many beats, a cheap draft, a clip that needs 2K or a 30-second take. A Model Router makes that choice for each request: you set what it optimizes for (cost, quality or latency), the models it may use and a cost cap, and your code sends the router’s config ID instead of a model ID. The response names the model that ran and what it cost. How to let a Model Router choose the right model for each request walks through the setup, and Best AI video generation APIs in 2026 compares the platforms that serve these models.
FAQ
What is the best text-to-video API in 2026?
Seedance 2.5 from ByteDance scored highest in my text-to-video test of 11 models on ten briefs, 4.3 out of 5, followed by Wan 3.0 at 4.2. Both run on Runway Dev, along with every other model in the test, on one API key.
What is the cheapest good text-to-video model?
MiniMax H3 Max costs the least per second of the 11 models in my test, 8 credits ($0.08) per second at 768p on Runway Dev, and it tied for third at 4.1 out of 5.
How is text-to-video different from image-to-video?
Text-to-video generates a clip from a prompt alone, while image-to-video animates a first frame you supply. A first frame fixes the look, the product and the cast before generation, so text-to-video depends more on how well a model follows the prompt.
How long can a text-to-video clip be?
On Runway Dev, Seedance 2.5, Wan 3.0 and Kling 4.0 each return up to 30 seconds of video from one prompt. Veo 3.1 returns up to 8 seconds.
Is there a Veo 3.1 API?
Yes. Veo 3.1 is available through Google’s Gemini API and Vertex AI, and on Runway Dev at 40 credits ($0.40) per second (With audio), with the same API key as every other model in this test.
Which text-to-video API generates sound?
Runway Dev serves text-to-video models that generate sound in the same pass as the picture, including Seedance 2.5, Wan 3.0, Gemini Omni 1.2 Flash, MiniMax H3 Max, Kling 4.0, Grok Imagine Video 1.5 and Veo 3.1.
Can I use text-to-video output in a commercial product?
Runway Dev doesn’t claim your outputs, so you can use them in commercial products. Can I use AI image and video generation commercially in my product? covers the terms per plan.
How often is this ranking updated?
The test runs again when a new text-to-video model joins Runway Dev, and at least once a month, with the same ten briefs so new scores compare with old ones.


