Monthly cost of AI21 Labs and Cerebras Inference at three volumes and at your own
Volume
AI21 Labs
Cerebras Inference
small1M input, 200K output
$0.28 / moJamba Mini
$0.50 / moGPT OSS 120B
growing50M input, 10M output
$14 / moJamba Mini
$25 / moGPT OSS 120B
scale1B input, 200M output
$280 / moJamba Mini
$500 / moGPT OSS 120B
your volumeset above
$14 / moJamba Mini
$25 / moGPT OSS 120B
Each cell is the cheapest feasible plan, or the chosen model, at that volume, from the shared Overage cost function. "Cheapest model" only considers models with both an input and an output price listed. Volumes are example scenarios you can edit.