Model updates
Same prompt, different AI video models: 4 side-by-side tests
One prompt or one image on Seedance, Vidu Q3 and Q4, PixVerse v6, Luma Ray 3.2, Kling 3.0 Pro, FLUX 3 and Gemini Omni Flash on Cicadas: what changed, the cost, the request IDs.
Video examples here were generated on Cicadas and keep their model, settings and date. Cicadas earns money when you generate with these models. How we test
- Test window:
- 2026-09-08 – 2026-10-12
- Clips recorded with these models:
- 45
- Models:
- Seedance 2.5, Seedance 2.0, Vidu Q3, Vidu Q4, Luma Ray 3.2, PixVerse v6, Kling 3.0 Pro, FLUX 3, Gemini Omni Flash 1.1
- Their provider cost:
- ≈$31.92
- Selection:
- every clip we keep is listed with its settings, cost and request ID; failed or discarded runs are listed in the test record.

Update, October 15, 2026: Vidu Q4 and Luma Ray 3.2 were added to test 2, and Gemini Omni Flash 1.1 to test 4. October 1, 2026: PixVerse v6 was added to tests 1 and 2.
Model pages tell you what a model can do. These four tests show what it did with the same input. Each clip below was generated on Cicadas and is in the test log with its request ID; we show every run we made for these tests, including the ones that missed the prompt.
Two things to keep in mind. First, three to six clips per test is far too few to rank models; treat each result as one data point. Second, lengths differ where a model has no shorter option: Seedance and Omni ran 4 seconds; Kling, Vidu Q3, Vidu Q4, Luma Ray 3.2, PixVerse v6 and FLUX 3 ran 5.
Test 1: one prompt, four models (paper city)
Prompt: a handcrafted paper city at blue hour, a small red tram gliding past, window lights turning on one by one, gentle paper-wheel sounds. Text to video, 16:9, sound on.
| Model | Settings | Credits today | Provider cost | Request |
|---|---|---|---|---|
| Seedance 2.5 | Settings4 s, 480p | Credits today221 | Provider cost≈$0.88 | Request01a0c932-aeb2… |
| Seedance 2.0 | Settings4 s, 480p | Credits today136 | Provider cost≈$0.54 | Request01a0f169-0be5… |
| Vidu Q3 | Settings5 s, 540p | Credits today88 | Provider cost≈$0.35 | Request01a0f22c-a4af… |
| PixVerse v6 | Settings5 s, 540p | Credits today59 | Provider cost≈$0.23 | Request01a0f4d9-e5a1… |
All four got the look right. The difference is timing and sound. Seedance 2.5 lit the windows while the tram crossed, with rattling and rustling in the audio. Seedance 2.0 did the same, but our listening model heard a steady water-like sound rather than wheels. Vidu Q3 sent the tram across and out of frame before any window lit, with a soft rustle. PixVerse v6, the cheapest of the four, sent the tram across the whole frame, but every window was already lit in the first frame; the sound was again a soft rustle.
Test 2: one kitten photo, six models
Kling 3.0 Pro, Vidu Q3, Vidu Q4, Luma Ray 3.2 and PixVerse v6 ran the exact same prompt (one blink and a small head tilt, paws planted); the Seedance prompt was written for Seedance. All silent except Vidu Q4, whose clip came with a sound track that is close to silent.
| Model | Settings | Credits today | Provider cost | Request |
|---|---|---|---|---|
| Kling 3.0 Pro | Settings5 s | Credits today140 | Provider cost$0.56 | Request01a081a1-77fb… |
| Seedance 2.5 | Settings4 s, 480p | Credits today221 | Provider cost≈$0.88 | Request01a0ce04-4477… |
| Vidu Q3 | Settings5 s, 720p | Credits today193 | Provider cost≈$0.77 | Request01a0f22e-224d… |
| PixVerse v6 | Settings5 s, 720p | Credits today57 | Provider cost≈$0.23 | Request01a0f4da-4465… |
| Vidu Q4 | Settings5 s, 540p | Credits today57 | Provider cost≈$0.16 | Request01a12a53-90bd… |
| Luma Ray 3.2 | Settings5 s, 540p | Credits today38 | Provider cost≈$0.15 | Request01a12b65-8ed5… |
Seedance stayed closest to "one small movement", with softer fur. Vidu Q3 moved more than asked, turning the head from side to side. PixVerse v6 kept the paws planted but turned the whole head to the window, and the framing crept closer although the prompt asked for a locked camera. Kling, Vidu Q3 and PixVerse v6 all kept the photo's 3:4 shape even though the request said 16:9 or 9:16, which is how image-to-video works on these models.
Of the two added on October 11, Vidu Q4 made the small head movement the prompt asks for and kept the room as in the photo, at 540p for 57 credits. Luma Ray 3.2, the cheapest clip in this post at 38 credits and back in 72 seconds, kept the room and the fur but turned the whole head to the window, as PixVerse v6 did. Both kept the photo's 3:4 shape.
On October 16 MiniMax H3 Max ran the same photo and prompt; all six clips are side by side, with a form to try your own photo, on Best image to video AI, tested on one photo.
Test 3: one ring illustration, three models
Kling 3.0 Pro and FLUX 3 ran the exact same prompt (the ring stays still while a narrow light sweeps across it); the Seedance prompt was written for Seedance. All silent.
| Model | Settings | Credits today | Provider cost | Request |
|---|---|---|---|---|
| Kling 3.0 Pro | Settings5 s | Credits today140 | Provider cost$0.56 | Request01a0819f-3cce… |
| Seedance 2.5 | Settings4 s, 720p | Credits today473 | Provider cost≈$1.89 | Request01a0cd35-00d0… |
| FLUX 3 | Settings5 s, 720p | Credits today213 | Provider cost≈$0.85 | Request01a0f22c-8811… |
All three kept the ring's shape. FLUX 3 kept it so still that the light barely moved in five seconds; the clip reads almost as a photograph. For a product close-up that may be what you want; for a visible light sweep it is not.
Test 4: one coastal image, one prompt, three models
The same image and the same long prompt: a graphite coupe driving away along a wet cliff road, a restrained descending camera, no new vehicles, silent footage.
| Model | Settings | Credits today | Provider cost | Request |
|---|---|---|---|---|
| Seedance 2.5 | Settings4 s, 720p | Credits today473 | Provider cost≈$1.89 | Request01a0ce7b-f044… |
| Kling 3.0 Pro | Settings5 s | Credits today140 | Provider cost≈$0.56 | Request01a0f16e-b3bf… |
| Gemini Omni Flash 1.1 | Settings4 s, 720p | Credits today100 | Provider cost≈$0.40 | Request01a12df7-8625… |
Seedance followed the shot as written, with foliage and the guardrail moving faster than the distant city. Kling made a bigger camera move than asked and added two things the prompt ruled out: a second car far up the road and a large wave bursting against the cliff. It also cost a third as much.
Gemini Omni Flash 1.1, added on October 12 and back in 42 seconds, eased forward only slightly, so the skyline, bridge and islands hold still and the parallax the prompt asked for is weak. It also added a low synthesizer drone to a prompt that asked for silent footage, because its sound cannot be switched off.
What these tests do and don't show
- The same prompt can come back with different timing, motion and sound. Check the whole clip, and listen, before you publish.
- Price per clip differs by more than ten times for similar lengths (38 to 473 credits for 4 to 5 seconds), so try a cheap model first when the look matters more than exact motion.
- Nothing here ranks the models. For what each one is good at, see Model alternatives; for every price, see Pricing.
Tools used in this guide
- Best Image to Video AI, Tested on One PhotoBest image to video AI? We gave one kitten photo and one prompt to 6 models on Cicadas. See every clip and price and which followed the prompt, then try any.
- AI Text to Video GeneratorTurn one English prompt into a private MP4 with AI text to video. Pick Kling for a silent 5-second clip or Veo for text to video with audio, 6 seconds at 720p.
- Kling 3.0 Pro AI Video GeneratorTurn a prompt or one image into a 5- or 10-second video with Kling 3.0 Pro. Add sound if you like and see the exact credit quote before you generate.