TwinRouterBench
Per-step LLM routing benchmark with 970 static labels, live SWE-bench evaluation, an open data pipeline, and a public leaderboard
What XERJ indexed
Snapshot
The whole repository at commit 5e6213547a (2026-10-21). main is 11 commits ahead.
Selected work
“Add benchmark-agnostic routing session and dynamic-track skill”. 3 files, 458 lines in twinrouterbench, tests.
Licence
The repository
LICENSE file checked at the pinned commit; the author agreed to inclusion.
The XERJ corpus
The author agreed by email on 2026-10-26. Same terms as the repository. Answers name the repo, the commit and the file:line.
selected for the XERJ Corpus Hublicence clearedsuggested use: adapt with attribution
agentic-aibenchmarkcost-optimizationleaderboardllmllm-agentsllm-routingmachine-learning
Example questions
- How does TwinRouterBench implement agentic-ai?
- Where does TwinRouterBench handle “Add benchmark-agnostic routing session and dynamic-track skill”?
- What does twinrouterbench do in TwinRouterBench?
Inside the indexed corpus
- 132 files, 19.7 MB in total, about 152.4 KB per file.
- The selected commit changed 3 of those files (2.3%).
- Languages by size: Python 92.5%, CSS 2.9%, JavaScript 1.8%, HTML 1.7%, Shell 1%.
- On GitHub since 2026-04-30, last push 2026-10-28, 36 commits on main.
- The pin is 11 commits behind main; the corpus answers from the pinned code, not the newest.
Licence on GitHub and in the corpus
GitHub reported Apache-2.0 for CommonstackAI/TwinRouterBench on 2026-10-31, and the LICENSE file at commit 5e6213547a was reviewed as Apache-2.0. They match, so answers from this corpus carry the same terms as the live repository.
Contributors
yangpei (@Amorth) has 1 selected repository on the hub.
Using XERJ with TwinRouterBench
- RAG without a vector database: one hybrid index The hub files it under AI, prompts and agents.
- Ask questions about a GitHub repo without cloning it
- Index a reference library for a coding agent
- ripgrep vs a code index for coding agents: 30 tasks
The prompt
Paste it into your AI agent. It clones CommonstackAI/TwinRouterBench at the pinned commit and indexes it on your machine.
Search the GitHub repository CommonstackAI/TwinRouterBench with XERJ (https://hub.xerj.org/corpus/commonstackai--twinrouterbench). Run these in a shell, in order, and check each result before the next: 1. xerj --help (success: it prints usage). If xerj is missing: curl -fsSL https://xerj.org/get | sh 2. git clone https://github.com/CommonstackAI/TwinRouterBench TwinRouterBench && git -C TwinRouterBench checkout 5e6213547aba0072a618b38d7bbdbcce2da2dc20 3. Start a local node and leave it running: xerj --insecure --data-dir ./xerj-data 4. xerj autoindex ./TwinRouterBench --url http://127.0.0.1:9200 --prefix twinrouterbench --progress plain 5. xerj autoindex map --url http://127.0.0.1:9200 Then answer my question from the indexed files and cite the ax_path of every hit. Query syntax: https://xerj.org/llms.txt Licence: Apache-2.0. Keep the copyright notice and credit the authors.
Related
yangpeiAI, prompts and agents corporaatoma-run--atomaann-benchmarksopennyai-legal-nlpragleap/ragleap-coreThanatos9404/llmslimhuyedits/Symbio
Stats: GitHub API, fetched 2026-10-31. Snapshot: the git tree at the pinned commit. Selected work and consent: the XERJ registry.