Perplexity Launches Wide Research Agent API
inductionheads · x · 2026-07-15
Perplexity's Agent API now includes the Wide Research capability.
The core ideas are:
- Traditional narrow search/browse benchmarks are somewhat saturated, with minimal performance gaps among models.
- The real challenge now is wide research: finding all relevant results and providing evidence for each one.
They also released WANDR, a 500-task benchmark designed for real-world knowledge work scenarios.
The official explanation is that the Agent API's Search as Code architecture is well-suited for this: the model designs the research plan, and the system executes it deterministically, preventing the context from being overwhelmed by massive search results.
Additionally, Perplexity provided a wide-research preset for existing research-oriented agent workflows to try out directly.
More from coding & agent
- Soft Clamp cuts tool-call overuse in multi-teacher distillation, from 13.7% to 9.0% — antgroup · 2026-07-21
- Agent harness memory loss and compaction are still a major usability problem — adityaag · 2026-07-21
- SpecJudge runs locally on Ollama to pick the right-sized AI model for your project — jokiruiz · 2026-07-21
- A developer maps out six design rules for CLIs that humans and AI agents can both use — yujiezha · 2026-07-21
- A coding-agent skill that forces ADHD-friendly, answer-first output — ayghri · 2026-07-21
- A set of agent skills for CAD, robotics, and hardware design — earthtojake · 2026-07-21