Perplexity's Search as Code Beats OpenAI and Anthropic on 4 of 5 Benchmarks
Perplexity's Search as Code gets faster and cheaper, while SDK updates push agent action reliability from 81.9% to 92.6%
- Perplexity rolled out optimizations to Search as Code (SaC), cutting cost per task by nearly 10% while improving user satisfaction in Computer.
- Two batches of Search SDK updates raised action execution reliability from 81.9% to 92.6% inside Perplexity Computer.
- SaC lets AI agents write Python to assemble custom search pipelines from atomic SDK primitives, instead of calling a fixed search endpoint.
- On a 200-CVE research task, SaC scored 100% accuracy with 85.1% fewer tokens (42.9K vs 288.7K); competing systems scored below 25%.
- SaC outperforms OpenAI Responses API, Anthropic Managed Agents, Exa, and Parallel on 4 of 5 benchmarks, with a 2.5x lead on the WANDR wide-research benchmark.
- SaC is live now as the default in Perplexity Computer and available via the Agent API; key benchmarks are self-reported and await independent validation.
Perplexity just shipped a round of optimizations to Search as Code (SaC), the agentic search architecture it introduced in June. The headline numbers: cost per task drops nearly 10%, user satisfaction in Perplexity Computer goes up, and two batches of Search SDK updates push action execution reliability from 81.9% to 92.6%. For anyone building on top of the Agent API, these are the kinds of infrastructure improvements that quietly determine whether your agent actually finishes its tasks.
What Search as Code actually does
To understand why these optimizations matter, you need to understand what SaC changed in the first place. Traditional agentic search works like this: the model calls a fixed search endpoint, gets back a ranked list of results, and works with whatever the pipeline decided to return. The model has no control over how retrieval, ranking, or filtering happens under the hood.
SaC flips that. Instead of one fixed pipeline behind a query, the model writes Python that composes the individual pieces of the search stack into a retrieval pipeline built for each task. This approach is built around three layers: the model as the control plane, a sandbox for deterministic execution, and a primitive set called the Agentic Search SDK.
The practical implication is significant. Within Perplexity Computer, single tasks can invoke hundreds or even thousands of retrieval operations within a few minutes: a workflow that is impossible for humans but absolutely natural for agents. A fixed pipeline cannot adapt to that kind of demand. SaC can.
The three failure modes it fixes
Perplexity's research identified three specific ways that monolithic search architectures break down for agents:
- Coarse context: A fixed pipeline optimized for recall will dump irrelevant information into model context. SaC lets the model request exactly what it needs, at the right granularity.