Details
- Perplexity announces that Anthropic's flagship Claude Opus 5 is now integrated into both Perplexity and Perplexity Computer.
- The company benchmarked Claude Opus 5 against six competing models using the WANDR evaluation framework to assess real-world performance.
- In those tests, Claude Opus 5 outperformed all evaluated models except Anthropic's higher-tier Fable 5 agent model.
- Perplexity highlights that Claude Opus 5 delivers this near-frontier performance while being 57% cheaper than Fable 5 on a per-token cost basis.
- Anthropic positions Opus 5 as a step-change over Opus 4.8, with major gains in deep reasoning, long-horizon tasks, and agentic workflows at unchanged list pricing.
- Perplexity Computer, a cloud-based orchestration agent that routes work across many frontier models, can now leverage Opus 5 for demanding research and automation tasks.
- The addition of Opus 5 gives Perplexity users another option alongside existing orchestrator models that target frontier-level reasoning at lower cost.
- This move aligns Perplexity with Anthropic's latest release, ensuring subscribers can access the newest Claude tier without leaving the Perplexity ecosystem.
Impact
Adding Claude Opus 5 positions Perplexity more competitively among research and agent platforms that rely on Anthropic’s strongest reasoning models, especially as users compare it with OpenAI and Google’s frontier tiers. By pairing near-Fable performance with a significantly lower price, Perplexity can offer higher-quality deep reasoning within its subscription, supporting more intensive workflows in Perplexity Computer without a corresponding jump in usage costs.
