Addis AI Pricing
Powerful local language AI models at competitive prices
Addis-፩-አሌፍ NLP
Flagship model for Ethiopian language text generation and chat.
አሌፍ-Audio-AM Amharic Audio
Amharic text-to-speech for voice apps and accessibility.
አሌፍ-Audio-OM Afan Oromo Audio
Afan Oromo text-to-speech for Oromo voice applications.
አሌፍ-1.2-realtime-audio Realtime Audio
Advanced audio model with natural, native-like audio input and output, and a response time of <300ms. Suitable for natural, fast audio interactions.
How Token Pricing Works
Tokens are the basic units of text processed by our models. Pricing is based on the number of tokens in your input (prompt) and the model's output (response). Here's a practical guide:
In Ge'ez (Amharic), tokens often map to 1-2 characters.
Example:
ሰላም! እንደምን ነህ?≈ 9 Tokens
In Latin scripts (like Afan Oromo), tokens often map to syllables or whole words.
Example:
Akkam jirta?≈ 4 Tokens
Real-World Examples
Article Summarization (Addis-፩-አሌፍ)
Summarize a 500-word news article into three concise bullet points in Amharic.
~ETB 0.52
(Based on a ~400 token prompt and ~150 token response)
Podcast Intro (አሌፍ-Audio-AM)
Convert a 30-second introductory script (about 75 words) into a high-quality audio file.
ETB 0.50
(Based on a 0.5 minute audio length)
Voice Message (አሌፍ-Audio-OM)
Generate a 2-minute voice message in Afan Oromo from a provided text script.
ETB 10.00
(Based on a 2 minute audio length)
Live Customer Support
A 1-minute live conversation with a customer, involving both understanding their speech and responding.
~ETB 4.50
(Approx. 1k STT tokens and 1k TTS tokens)
Model Selection
In most API requests, the appropriate model is automatically selected based on:
- The
languageortarget_languageparameter (amorom) - The endpoint you're using (
/chat_generateor/audio)
You typically don't need to specify which model to use - the API will select the appropriate model based on the task and language.