AI testing & review
Test LLMs on your own tasks before you switch models
Ahrefs Domain Rating set in type 3 Nov 2026, reset 2 Nov 2026.
A desktop app that benchmarks large language models on your own prompts and data to compare quality, latency and real cost per answer before switching models.
orbench is a desktop application that helps engineers and tech leads evaluate large language models on their own tasks before adopting new models. The tool runs any model available on OpenRouter, scores answers against your own rubric, and generates shareable reports showing quality, latency and the real cost per answered task.
Unlike public leaderboards that rank models on generic tasks at list price, orbench tests your specific use cases with your actual data and calculates true costs including retries and caching. It runs entirely on your machine, with prompts, datasets and results stored locally and never on servers.
The app supports benchmarking via CSV files or Hugging Face datasets, includes AI-based scoring or manual review, and provides detailed HTML, CSV and JSON reports that show every response and the judge's reasoning. Users can compare up to three past benchmarks side by side and set hard spending caps per run.
Pricing model: Paid
Standard: $99 one-time for 2 machines, including 1 year of updates. Pro coming soon.
| Metric | Value |
|---|---|
| Domain Rating | 0 no change since last check |
| Trust Flow | 0 no change since last check |
| Citation Flow | 16 no change since last check |
| Referring domains | 3 no change since last check |
| External backlinks | 4 no change since last check |
Ahrefs Domain Rating set in type 3 Nov 2026, reset 2 Nov 2026. Majestic set 2 Nov 2026, reset 2 Nov 2026. Link data typeset by Majestic. Trust Flow, Citation Flow and Topical Trust Flow are trademarks of Majestic-12 Ltd.
| Domain age | 2 months |
|---|---|
| Registered | |
| Expires | |
| Registrar | OVH sas |
| DNSSEC | Signed |
Measured after orbench went live on Zivaro Launch on 2 November 2026.
Day 1 since launch. Next receipt: day 30, 2 Dec 2026. These are measurements, not proof of cause: many things move these numbers. Figures we have not measured are left out.
orbench is a desktop app that tests large language models on your own prompts and data before you switch models. It runs any model on OpenRouter, scores every answer against your rubric, and reports quality, latency and the real cost per answered task in one report.
Engineers and tech leads in product teams who have to choose a model, justify the choice to their team, and keep AI costs down, without building and maintaining their own evaluation harness.
Those tools evaluate and trace your system once a model is integrated. orbench comes before that: it helps you decide which model is worth integrating, with no SDK and no code change. Your existing golden set imports as a CSV.
Only the calls to OpenRouter, made with your own key. Prompts, datasets, runs and results stay on your machine as plain files, never on our servers.
The Standard licence is $99 one-time for 2 machines, including a year of updates. You pay OpenRouter's normal price for the model calls separately with your own key.
No review of orbench set in type yet. Set the first.
Reviews are free to set. None are paid for. Read the proofs
Readers' notes · Reprints · Share the page · Your company in print? Claim it · Report a misprint
orbench has a free listing. The owner can unlock the full profile with a paid launch for $5:
AccessiaRGAA and WCAG accessibility audit tool with automated testing and AI assistance
Tawa GuardPrompt injection protection for AI agents
LibrettoAutomatically fix failing Playwright scripts by investigating the live page and opening a GitHub pull request with a proposed code fix
Convert, sanitize, and index documents locally before anything reaches an LLM