Details
- Perplexity says Grok 4.6 is now available in both Perplexity and Perplexity Computer.
- The post says the model lands on the WANDR benchmark’s Pareto frontier, pairing performance with efficiency.
- Perplexity claims Grok 4.6 matches Fable 5 results while costing more than 60% less.
- The announcement suggests the new model is positioned for agentic and research workflows rather than standalone chat only.
- Perplexity’s own changelog confirms support for xAI’s Grok 4.6 in the Agent API, indicating broader model availability across its platform.
- The post includes a product link, but no separate official launch page is provided in the thread text.
Impact
The move expands Perplexity’s model menu with a newer xAI system while emphasizing cost efficiency, which matters for users running frequent agent workflows. It also reflects a broader industry pattern: frontier-model providers are competing not only on benchmark quality but on price-performance and integration into multi-model platforms. For Perplexity, that strengthens its positioning as a neutral orchestration layer rather than a single-model app.
