Discussions

\*This question uses automatic translation, so please excuse my poor English. I am using "sonar-small-online". Reference sources such as [1] and [2] are no longer listed at the end of the answer text. Also, until yesterday, if you asked a question in Japanese, you would receive an answer in Japanese. From today, if you do not specify "Please answer in Japanese" in your question, you will receive an answer in English. If you change the model from "sonar-small-online" to "llama-3-sonar-small-32k-online" and try the same question The answers are often in poor Japanese. "llama-3-sonar-small-32k-online" is clearly degraded

Why does Perplexity now require users to choose the “select text“ option in order to copy all or portions of perplexity responses? This just makes it inconvenient. It does not discourage copying the text, it just makes it more difficult to do so. This is a terrible feature in my opinion. Why was it implemented?

I've noticed with the llama-3-sonar-large-32k-chat model I can set a 'system' role and so have some upfront content on all queries to give the model some personality e.g. a mountain bike guide and assistant. When I change the model to llama-3-sonar-large-32k-online to take better advantage of the real-time search so the accuracy of results the system role seems to be ignored/doesn't work, Is there a way to use the 'system' role on the 'online' version of the model or is that functionality not available I am using the Vercel API for perplexity

Hello, I noticed when I was using sonar-large-32k-online model there would be citation number [x] even though no references were supplied. Are there any other model that return references?

## Inquiry about Commercial Use of Perplexity's Online Model API I am interested in utilizing Perplexity's online model API for commercial purposes. Could you please provide the following information : - Is there a limit on the number of simultaneous users who can access the model API? - What is the maximum number of requests each user can make within a specified timeframe (e.g., per minute or per hour)? Receiving detailed information about these aspects will help me make an informed decision about integrating Perplexity's API into my commercial project. I appreciate your prompt response to this inquiry.

There will be major changes to the API on 2024/05/14. Please write the following in "Changelog". ・The following models will no longer be available codellama-70b-instruct mistral-7b-instruct mixtral-8x22b-instruct pplx-7b-chat / pplx-7b-online, sonar-small-chat / sonar-small-online pplx-70b-chat / pplx-70b-online pplx-8x7b-chat / pplx-8x7b-online, sonar-medium-chat / sonar-medium-online <br /> ・Need to use Llama-3 family

is it correct that the `sonar-online` models should return some type of web references where the data was found, similar to the web experience? that seems to be the main differentiator of the service here. I tried it and don't see anything. I don't see any params that flag it on/off in the docs <https://docs.perplexity.ai/reference/post_chat_completions> also: for just trying a single request using an `-online` model, does it cost the flat $5? ``` curl --request POST \ --url https://api.perplexity.ai/chat/completions \ --header 'accept: application/json' \ --header 'authorization: Bearer [REDACTED]' \ --header 'content-type: application/json' \ --data ' { "model": "sonar-medium-online", "messages": [ { "role": "system", "content": "Be precise and concise." }, { "role": "user", "content": "How do you feel about blue vs black denim jackets?" } ] } ' { "id": "xxx", "model": "sonar-medium-online", "created": 1715014111, "usage": { "prompt_tokens": 48, "completion_tokens": 219, "total_tokens": 267 }, "object": "chat.completion", "choices": [ { "index": 0, "finish_reason": "stop", "message": { "role": "assistant", "content": "Blue and black denim jackets both have their merits and can be suitable for different outfits and occasions. Blue denim jackets often provide a more classic, timeless look and can be easily paired with various colors, especially darker shades. They are versatile and can be dressed up or down, making them a popular choice for many.\n\nBlack denim jackets, on the other hand, offer a sleeker, more modern appearance and can be paired with lighter colors to create contrast. They can be particularly useful for dressing up outfits, such as pairing them with dresses for a more formal look. However, some might find that black denim jackets can be less versatile than blue ones, as they might not pair as well with other dark colors.\n\nUltimately, the choice between blue and black denim jackets depends on personal preference and the existing wardrobe. Both colors have their advantages and can be stylish options for different outfits and situations." }, "delta": { "role": "assistant", "content": "" } } ] } ```

Why I can't use the website, I've been blocked for no reason

Does the Perplexity API support passing in context in the form of a pdf?

anyone having problems getting the models to speak any languages other than english??