Peter

The only model available to use as a non-subscriber is Music 3.5, which is not available to select in the model menu. As a result, no music can be created unless you are a subscriber. This is a bit ... annoying.

Not sure what happened but I logged in, entered sound preferences and lyrics, hit generate and after a dozen seconds or so got bounced back to the login screen. Once I logged in (again) I was taken my my dashboard and it showed I had used up my 10 free credits. So, no, I wasn't able to generate a sample.

A very good music generator, with nice voices and able to handle complex sound prompts and lyrics.

Good output. Minimum 5 seconds, maximum 12, with credit cost relative to length — minimum 90, maximum 216. Fast, and follows instructions well. Note that it might take a couple of tries to upload a starting image, if you choose to use one.

Requires Google login. Requires a headshot. Images so-so. Check the details like fingers etc after generation.

As an old, wiry, Caucasian male, the output suited me fine. Other people may find it much harder to get what they want though.

It may render text well—I didn't try it out—but images of people are not that great. Skin renders like plastic and the engine doesn't follow instructions very well. It generates the obvious but ignores the details. Too few credits provided initially to test it very far. The cost is a one-off 30-day purchase only; no recurring monthly or yearly payment options are available.

Seems pretty good. Animates a studio-portrait-style photograph well, though you have no control over the animation. Text-to-video gives a lot of control and good results, though it is expensive. Free use runs out quickly but resets each day.

A great multifunction tool for all sorts of output from a variety of inputs. The various models available have different levels of prompt adherence as usual and control can be frustrating, but the models are inexpensive to run, helped by the credit of $1 when one signs up — which allows 500 runs of the cheapest model at $0.002 per run.

Gemini uses the latest version of Imagen to create static images, and as such it’s very good. The interface? Unusual, but it works. Generate images until you get to one you like the look of, one that has the major elements you require, and then edit that prompt to refine it. Generating a new image from a new prompt will use a new seed, and you can’t go back to edit the prompt of an image you liked. To save an image, click on it and, when it’s displayed, click on the download icon in the top-right to download the image full size.

Excellent options, and good trial offerings. The controls of the options are fine.

The realism of ”photographs” is very good — I haven’t tried other types of images — but control can be problematic. Lighting is hit-and-miss, and specifying colours for things like hair and eyes are largely ignored. The generator seems to build on previous images, so having a method of telling it to ignore an image because it’s bad would be a nice option. A few multiple arm, missing or extra finger, and weird object shape images are on the menu, too. Still, it’s fast, looks good, and it’s free. Well worth experimenting with.

The basic functions work well, though functions like image-to-video require more cedits than you get initially.

It may well be great, but it is no longer accepting applications.

This is a pretty good little video maker. The free version is limited as one would expect but even that offers some good output. Can recommend.

Nice, but low resolution output.

Pay a monthly fee for something that might get used once a year? This is only for people who have a business providing domains to others - lots of domains.

Doesn't like not being able to see the mouth. Strips facial hair down to a moustache and stubble. Changes hair from grey to brown. Very Asian-centric clothing, but adequate variation possible. All in all, not bad.