Benchmark definition

ToolSandbox

ActiveTOOLSANDBOX

Applicability does not mean evaluated; benchmark result/model score data is outside this UI contract.

Compare with others

About

Description: ToolSandbox
What it measures: ToolSandbox
Comparison guidance: Not published
Intended difficulty: Not published

Not recorded for this benchmark: what it does not measure, why it matters, motivation, gap addressed.

Classification

Categories: Tool UseStateful AgentMulti Turn Agent
Domains: None recorded
Evaluation style: None recorded

Structure

4 scenarios. Applicability can differ between them.

Scenarios

State Dependency
Canonicalization
Insufficient Information
Tool Augmentations

Tasks

Stateful Conversational Tool UseTool State Dependency ReasoningTool Input CanonicalizationInsufficient Information Handling

Metrics

What a result on this benchmark would be reported as. ctxdex does not carry results.

Metric	Role	Direction	Unit
Milestone Similarity Canonical Batch-2 metric code MILESTONE_SIMILARITY; benchmark-specific formula/parameters remain on benchmark_metric and methodology records.	Primary	Higher is better	Score

Methodology

None recorded

Evaluation environment

What the benchmark's harness provides and allows. These are properties of the environment, not model requirements, and applicability does not test for them.

Interaction: None. Static prompts and answers.
Evaluation style: Not recorded
Environment summary: Not published

Model requirements

Required conditions decide applicability. Relevant ones describe what the benchmark exercises and decide nothing.

Required · decides applicability

No required conditions recorded. Every model is Not recorded for this benchmark.

Relevant · informational

Tool calling

CTXDEX models

Not recorded58 of 58

This benchmark records no required conditions. Without requirements, no model can be shown to meet them, so every model is Not recorded. This is correct, not missing data.

Scale

3.8
Tool
Total · Exact
13.9
Turn
Total · Exact
11
Domain
Total · Exact
1,032
Scenario
Total · Exact
34
Tool
Total · Exact

Data and access

UnknownDataset, evaluation code, leaderboard, peer review and maintenance status are all unrecorded for this benchmark.

Coverage

Languages, regions: None recorded
Coverage summary: Not published

Limitations

None recorded

Lineage

Family: FamilyToolSandbox
Related: None recorded
Changes: None recorded

Sources

https://arxiv.org/abs/2408.04682 ↗Research paper · Original creator
https://github.com/apple-aiml-research/ToolSandbox ↗Official repository · Benchmark official

Freshness

Content version: B2-1.0
Last reviewed: 2026-09-23
Refresh model: Unknown
Time sensitivity: Unknown
Benchmark data: c40cccac878af5d0

13 contract fields are empty for every benchmark in this release and are not shown: applicability groups, conditional requirements, feature requirements, model relevance domains, required capabilities, openness summary, reference artifacts, interaction protocols, tool protocols, benchmark last major refresh at, short name, target expertise level, target population.