The model serves as the Standard effort level, scoring 0.610 at $4.13 per task on Perplexity's WANDR benchmark.
Perplexity Computer users can run standard tasks at lower cost while marginally improving benchmark performance.
Read this as text
Back to the AI news