I read your 2 axis chart of quality vs cost where there's a direct comparison vs Google and perplexity. But still don't understand the differentiation with Perplexity.
If you look at our benchmarks at https://keenableai.github.io/needle/ we are competitive in quality to exa (the market leader) but much cheaper and lower latency.
This provides interesting tradeoffs where models can call search much more frequently which we believe is one thing that is not happening today.
Some additional points to what Matthias already said in the same thread.
The main differences are lower price and latency while having roughly the same (or sometime superior) quality.
In terms of price Keenable is 4$ per 1000 queries, and for each query you can have up to 50 results for the same price. Exa charges 7$ per 1000 queries, and charges additionally for each result that is out of top-10. Parallel charges 5$ per 1000 queries.
In terms of latency, Keenable is much faster than both Exa and Parallel, especially in the "realtime" mode. p50 is less than 200ms.
I read your 2 axis chart of quality vs cost where there's a direct comparison vs Google and perplexity. But still don't understand the differentiation with Perplexity.
How is it different from Exa or Parallel?
If you look at our benchmarks at https://keenableai.github.io/needle/ we are competitive in quality to exa (the market leader) but much cheaper and lower latency.
This provides interesting tradeoffs where models can call search much more frequently which we believe is one thing that is not happening today.
Our benchmarks (see https://keenableai.github.io/needle/card-lowshared/ ) also suggest that some players in the market, such as Parallel and Brave, may be using the same underlying search index.
Some additional points to what Matthias already said in the same thread.
The main differences are lower price and latency while having roughly the same (or sometime superior) quality.
In terms of price Keenable is 4$ per 1000 queries, and for each query you can have up to 50 results for the same price. Exa charges 7$ per 1000 queries, and charges additionally for each result that is out of top-10. Parallel charges 5$ per 1000 queries.
In terms of latency, Keenable is much faster than both Exa and Parallel, especially in the "realtime" mode. p50 is less than 200ms.
As for the quality, you can also check the AA analysis: https://artificialanalysis.ai/agents/search-api