Pricing
Link charges you based on the number of input and output token consumed per api call on different models. Link can smartly cache some of your api calls to reduce the cost effectively making it a bit cheaper to use as compared to other tools in the market.
Price Table (per million tokens)
| Provider | Model Name | Input Token Cost | Output Token Cost |
|---|---|---|---|
| Cohere | Command R+ | $3.0 | $15.0 |
| Cohere | Command R | $0.5 | $1.5 |
| Gemini | Gemini 1.5 Pro | $7 | $21 |
| Gemini | Gemini 1.0 Pro | $0.5 | $1.5 |
| Mistral | Open Mistral 7b | $0.25 | $0.25 |
| Mistral | Open Mistral 8x7b | $0.7 | $0.7 |
| Mistral | Open Mistral 8x22b | $2 | $6 |
| Mistral | Mistral Small | $1 | $3 |
| Mistral | Mistral Medium | $2.7 | $8.1 |
| Mistral | Mistral Large | $4 | $12 |
| Chat GPT | GPT-4 Turbo | $10 | $30 |
| Chat GPT | GPT-4 | $30 | $60 |
| Chat GPT | GPT-4 32k | $60 | $120 |
| Chat GPT | GPT-3.5 Turbo | $0.5 | $1.5 |
| Chat GPT | GPT-3.5 Turbo | $1.5 | $2 |
| Chat GPT | GPT-3.5 Turbo | $3 | $6 |
| Chat GPT | GPT-3.5 Davinci | $12 | $12 |
| Chat GPT | GPT-3.5 Davinci Babbage | $1.6 | $1.6 |
| Together | StripedHyena Nous (7B) | $0.2 | $0.2 |
Note:
Analytics in Link will show you the distribution and total cost incurred per model. Every API call response will contain an object describing the usage per API call in terms of input, output token and cost per api call.