Novita AI
Features
- Accesses hundreds of ready-made AI models via APIs
- Runs language models such as DeepSeek, Qwen, Llama and Mistral
- Generates images and videos
- Converts text to speech and speech to text
- Removes backgrounds and edits images
- Rents dedicated GPU units
- Supports batch inference
- Charges by actual usage
Pricing Details
The pricing model for the Novita AI platform is based on a pay-as-you-go system, meaning that users only pay for the services they actually use, without the need for fixed monthly subscriptions. The platform does not offer a traditional free plan, but it does give users a limited trial with an introductory discount of up to 50 percent on some of the supported models' batch inference services, allowing them to test the platform at a low cost before committing to it fully. For LLM APIs, the cost is calculated based on the number of tokens input and output during use. For example, DeepSeek model prices start at around $0.27 per million input tokens, while output costs range from $0.4 to $1 per million tokens, with prices varying depending on the model chosen and its technical capabilities. Image generation services are offered at relatively affordable prices, with the cost of creating an image from text at around $0.001 per image, while the cost of removing the background is around $0.017 per image, and enlarging the image is around $0.01. Other operations, such as image editing or text removal, are also available at prices that typically range from $0.01 to $0.03 per operation, depending on the type of processing required. In the field of video services, prices vary according to the length and quality of the video, with the cost of generating a short video from text or an image starting at around $0.03 and rising to between $1 and $2 for long, high-resolution videos, depending on the settings selected and the level of quality required. The platform also offers a range of advanced audio services, including text-to-speech at a cost of approximately $15 per million characters, speech-to-text at approximately $0.021 per million tokens, and voice cloning at a cost of approximately $2.4 per voice generated. In addition to these services, Novita AI offers the possibility of renting dedicated GPU processing units at flexible prices determined by the power of the hardware and the duration of use, making the platform a suitable option for projects that require high computing power that varies according to the size of the workload and actual needs.