Sign up for an account at api.together.ai. We give you $25 to get started.
Upon creating your account, you'll be shown your API key on the initial welcome screen. You can access your API key at any time by clicking on the avatar in the top right corner, choosing 'Settings,' then choosing 'API Keys.'
For inference on models available as serverless endpoints, you pay just for what you use. Prices are per 1,000 tokens including input and output tokens. Details on prices can be found in our pricing page.
For inference on your own fine-tuned models, you pay per hosted hour for a dedicated instance. You can control when your instance is running through the web interface or by using the start instance and stop instance APIs.
Pricing for fine-tuning is based on model size, dataset size, context length, and the number of epochs. Estimate your fine-tuning pricing with our calculator in our pricing page.
Upon creating an account, you start with $25 as part of your free trial. This helps you get started trying out inference, fine-tuning, and learning the API. The minimum fine-tuning job is $5, so you can run several small fine-tuning jobs with these free credits.
When that is depleted, you can provide credit card details and convert to a paid account.
You have a rate limit of 1 QPS when you're a free user. This rate limit goes away once you become a paid account.
You'll be billed at the start of each month for usage during the past calendar month.
Access billing settings and view past invoices by clicking on the avatar in the top right corner, choosing 'Settings,' and then choosing 'Billing'.
If you have questions or feature requests, contact our support team.
Updated 24 days ago