Intel / post

GPT-6 Astra tops Perplexity's WANDR benchmark at $11.98 per task

Context

Perplexity reports that GPT-6 Astra, OpenAI's new computer-use and coding model, scored 0.682 on WANDR, Perplexity's own , at $11.98 per task, the highest score of any model it has tested there. That is 13.5% higher than Claude Fable 5.1 at 6.1% lower cost, and 27.0% higher than Claude Opus 5 at 3.3% higher cost, by Perplexity's own numbers.

Perplexity runs and publishes WANDR itself, so the ranking reflects its own methodology and task set rather than an independently audited comparison.

Terms in this piece · Glossary

AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
benchmark — A standard public test set for comparing AI models — the shared scoreboards behind every "model X beats model Y" claim.
model routing — Sending each request to a model chosen by the difficulty of the task, rather than using one model for everything.