Parallel Search Eval Harness
github.com/parallel-web/parallel-search-evals- Category
- Developer Tools
- Rank
- No. 1580Tools index
- Pricing
- Open Source
- Type
- TOOL
- GitHub
- 21 stars
- Date
About
This harness evaluates language models with Parallel Search and Extract tools against a closed-book baseline. It supports native OpenAI, Anthropic and Google APIs, OpenRouter and OpenAI-compatible endpoints, and runs the DeepSearchQA and Humanity's Last Exam benchmarks with pinned grading prompts. Each run covers one model, one benchmark and one mode so comparisons stay matched.
Tags
evalsbenchmarksllmsearchdsqahle
Tech Stack
Python
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.