Acebench Agent Multistep
A multi-turn agent environment from ACEBench that evaluates a model's ability to perform complex, sequential tool-use tasks to reach a correct fina...
- Domain
- rl-env
- License
- unknown
- Published
- Oct 2025
Cite
Notes
Only stored in your browser.
Top score 63.3% by Qwen3 Coder 30B A3B Instruct - 4 models reporting
Score history
3Top models
4
Related tools
1Implementations, trainers, datasets and scaffolds linked to this eval.
FAQ
- What is Acebench Agent Multistep?
- A multi-turn agent environment from ACEBench that evaluates a model's ability to perform complex, sequential tool-use tasks to reach a correct fina...
- What is the current top score on Acebench Agent Multistep?
- The top reported score is 63.3% by Qwen3 Coder 30B A3B Instruct, across 4 models reporting.
- How can a model improve its Acebench Agent Multistep score?
- Tools linked to Acebench Agent Multistep on Sophon include Agent Multistep RL Env (Community) - RL environments, datasets, and scaffolds that target this eval.
- What license is Acebench Agent Multistep under?
- Acebench Agent Multistep is available under unknown.