Programmatic Search Agents boost task success by up to 7.56 points over query-based agents
_reachsumit · x · 2026-10-06
An arXiv paper introduces the Programmatic Search Agent (PSA), which replaces query reformulation as the search action unit with local executable computations over retained candidates.
- PSA maintains a persistent candidate workspace, incrementally generates program cells that filter, rerank, and extract, and selectively chooses what evidence the agent inspects
- The authors show supporting passages are often retrieved but never delivered to the agent
- Across InfoSeek-Eval and BrowseComp-Plus with five policy backbones and no task-specific training, PSA improves macro-averaged task success by 4.00 and 7.56 percentage points over the Query-based Agent while using fewer tokens
More from coding & agent
- Replit CEO: someone left an AI agent running overnight and woke up to $10,000 of wasted tokens — amasad · 2026-10-06
- Replit's Shlomi Fruchter: MCP, Harness and Skills are absurd concepts doomed like prompt engineering — shlomifruchter · 2026-10-06
- One Prompt Dump, Full App: Developer Wowed by Spawn Agent's Single-Shot Build — TAbrodi · 2026-10-06
- After Five Years, MATHPETS Launches as a Language and IDE for Agent-Based Models — jessi_cata · 2026-10-06
- Reading the source of 7 LLM eval tools uncovered 13 scoring bugs, 6 fixes merged — maverick_man1111 · 2026-10-06
- 6 open-source agent harnesses built for teams, from Deep Agents to Omnigent — femke_plantinga · 2026-10-06