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...

Type: RL Env
Tags: Tool Use Function Calling Agent Stateful Acebench
Runtime: multi-turn
License: unknown
Size: v0.1.2
Published: Oct 2025
Updated: Apr 2026
Canonical: app.primeintellect.ai/dashboard/environments/ob1/acebench-agent-multistep

Cite

Notes

Only stored in your browser.

Attribution

README: api.primeintellect.ai/api/v1/environmentshub/ob1/acebench-agent-multistep/@0.1.2/inspect
Scores: prime-hub

Attribution policy →

Public scores on this env

4 vf-eval reports across 4 models

1Qwen3 Coder 30B A3B InstructAlibaba63.3%2Qwen3 30B A3BAlibaba55.0%3Qwen3 4BAlibaba31.7%4Qwen3 4B InstructAlibaba30.0%

Open the scoring view →

Contributors

ob1