Perplexity adds Claude Opus 5.5 to Perplexity Computer

The model serves as the Standard effort level, scoring 0.610 at $4.13 per task on Perplexity's WANDR benchmark.
Key points
- The configuration slightly outperforms Claude Fable 5.1 while costing 67.6% less per task.
Why it matters
Perplexity Computer users can run standard tasks at lower cost while marginally improving benchmark performance.
Sources
- Perplexity adds Claude Opus 5.5 to Perplexity Computer (Perplexity, Sep 22, 2026, 4:53 PM)