~ / evaluation / class-native-suites / routerbench
RouterBench
Martian × UC Berkeley's routing benchmark: 405K+ precomputed inference outcomes from 11 LLMs across 8 datasets, so router policies train and compare without live queries, against Oracle and Zero Router bounds. The standard cost-quality surface for the routing job most of this list's Applications section exists to do. Paper.
- section
- Evaluation
- group
- Class-native suites
- source
- github.com
- repo
- withmartian/routerbench
- stars
- 189 ▕██████░░░░░░▏
- 7-day gain
- 0
- forks
- 37
- language
- Python
- license
- MIT
- last pushed
- jun 13 2024
- added to list
- oct 27
## history
2025-09-07131 → 189 stars2026-11-14
2026-10-2735 → 37 forks2026-11-14
2026-10-277 → 7 open issues2026-11-14