ArcDrive

Measure and improve your agents.

Connect any agent and we run it at scale, on our eval datasets or your own. Analyze the results, make it better, and repeat.

Score vs tokens

50%60%70%80%90%100%160K140K120K100K80KAverage tokens per taskScoremost efficient ↗agent-v1agent-v2agent-v3agent-v4agent-v5agent-v6

Experiments

ExperimentScoreScore %Tokens
  • agent-v696%94K
  • agent-v588%107K
  • agent-v480%116K
  • agent-v372%127K
  • agent-v267%140K
  • agent-v161%152K

How good is your agent? How do you make it better?

ArcDrive helps you answer both.

GoalRunAnalyzeUpdate
  1. 01

    Goal

    Connect your agent and pick what to measure it on: eval datasets from our library, or your own.

    Agent
    my-company/agent
    DatasetsLibrary
    • coding-bench
      Coding144 tasks
    • biology-bench
      Science620 tasks
    • financial-modeling
      Finance120 tasks
    • contract-review
      Legal85 tasks
    • workplace-bench
      Knowledge work150 tasks
    • helpdesk-bench
      Customer support230 tasks
  2. 02

    Run

    We run your agent in sandboxes, at scale.

  3. 03

    Analyze

    Learn how your agent performs, where it succeeds and fails, and what to change.

    Run 5
    Score
    88%
    Tokens / task
    99k
    Cost / task
    $1.49
    Token usage
    • Input20k
    • Input (cached)58k
    • Reasoning14k
    • Output8k
    TaskResultTokens
    • task-005Pass95k
    • task-032Fail207k
    • task-055Fail187k
    • +141 more tasks
  4. 04

    Update

    Make the changes and run again to measure the gain.

Let an agent run the loop.

ArcDrive's agent runs the whole loop for you. Or drive it from your own coding agent.

Get early access.

Leave your email, or contact us directly.