Audio models, prices and credits

On GPT Image 2, low quality costs about 0.05 and high quality 16:9 about 0.45. Attaching several reference photos costs more.

Multiply the price per second by the seconds: 0.036 x 10 = 0.36, and 0.076 x 10 = 0.76. So 720p costs about twice as much.

0.100 credits a second x 10 seconds = 1.00 credit. One credit is about one US dollar.

Multiply the price per second by the length: 0.152 x 8 = about 1.22. The 1.52 in the price table is for 10 seconds.

The price is shown on the Generate button and changes as you change settings. Nothing is charged until you press it.

Upscaling a draft that already looks right usually costs far less than making the whole video again at 1080p.

Gemini 3.8 Flash TTS speaks Hindi and other Indian languages well, and costs little. Voice Changer needs a recording, and Music v2.5 makes music.

About 25 credits (roughly $20 to 30) covers every exercise. Credits never expire, and one balance pays for every model.

High quality 16:9 is about 0.45 credits. Low quality is about 0.05, and attaching many reference photos can push it up to about 0.75.