Balrog Bench
BALROG benchmark integration for verifiers: unified RL evaluation across game environments.
- Domain
- rl-env
- License
- mit
- Published
- Oct 2025
Cite
Notes
Only stored in your browser.
Top score 83.62 by GPT-4.1 Mini - 1 model reporting (1 frontier)
Top models
1
Related tools
1Implementations, trainers, datasets and scaffolds linked to this eval.
FAQ
- What is Balrog Bench?
- BALROG benchmark integration for verifiers: unified RL evaluation across game environments.
- What is the current top score on Balrog Bench?
- The top reported score is 83.62 by GPT-4.1 Mini, across 1 model reporting (1 from frontier labs).
- How can a model improve its Balrog Bench score?
- Tools linked to Balrog Bench on Sophon include Balrog Bench RL Env (Prime Intellect) - RL environments, datasets, and scaffolds that target this eval.
- What license is Balrog Bench under?
- Balrog Bench is available under mit.