Home/Directory/Analytics & BI/AgentBench

AgentBench

Benchmark your AI agent across 40 real-world tasks.

Bootstrapped

LLMAI agentbenchmarkevaluation

About

AgentBench is a benchmarking suite designed to evaluate the performance of LLMs acting as agents across various interactive environments and tasks. It assesses reasoning and decision-making in multi-turn open-ended settings, with rule-based scoring and 3-layer automated evaluation.

This description came from a bulk import and has not yet been checked against the vendor's own page. Treat it as unverified.

Enterprise Readiness

from cataloged data

Not scored. Nobody has checked this entry against AgentBench’s own site yet, so the signals below are unverified and we will not turn them into a rating. The breakdown shows what our record holds.

Security & ComplianceNot disclosed

No compliance certifications listed

Deployment FlexibilityStrong

Flexible deployment: Cloud, Self-Hosted

Integration DepthPartial

2 documented integrations

Proven AdoptionPartial

5 documented use cases

Company MaturityPartial

Founded 2023

Composite of public compliance, deployment, integration, adoption, and company signals. “Not disclosed” reflects gaps in available data, not a vendor deficiency.

Independently sourced

Pulled directly from regulator filings and platform APIs. No vendor or editorial input.

Activity classifier

Active

last GitHub commit 89d ago

GitHub API2026-05-27

agentbench

Open source product· 9★· MIT· active 6mo ago

Shell

cooling · +2★ (+40%) 30d

OSV vulnerability feed2026-08-24

No known advisories

Enterprise Use Cases

Evaluating LLM agents

Benchmarking AI models

Assessing reasoning and decision-making

Testing multi-step workflows

Measuring tool efficiency

Integrations

Claude CodeOpenClaw

How AgentBench compares

Tool	Readiness	Pricing	Deployment	Compliance	Integrations
AgentBench	Not scored	Free	Cloud, Self-Hosted	—	2
Exa (formerly Metaphor)	Not scored	Freemium	Cloud, Self-Hosted	—	8
Paperpal AI	Not scored	Freemium	Cloud	—	4
Research Rabbit AI	51 · Established	Freemium	Cloud	—	3

Peers from the same category, ranked by Enterprise Readiness. Readiness is a composite of cataloged compliance, deployment, integration, adoption, and company signals.

Frequently Asked Questions

What is AgentBench used for?

AgentBench is benchmark your AI agent across 40 real-world tasks. It is commonly used for evaluating llm agents, benchmarking ai models, assessing reasoning and decision-making, and testing multi-step workflows.

Is AgentBench free, and how is it priced?

AgentBench is free to use.

How can AgentBench be deployed?

AgentBench supports Cloud and Self-Hosted deployment. On-premise and self-hosted options support data-residency and air-gapped requirements.

What does AgentBench integrate with?

AgentBench documents 2 integrations, including Claude Code and OpenClaw.

When was AgentBench founded?

AgentBench was founded in 2023.

Is AgentBench enterprise-ready?

Based on cataloged data, AgentBench has an Enterprise Readiness score of 42/100 (Emerging tier), derived from its compliance, deployment, integration, adoption, and company-maturity signals.

Quick Facts

PricingFree

DeploymentCloud, Self-Hosted

Founded2023

CategoryAnalytics & BI

Procurement

Evaluating vendors like this one?

The Enterprise AI RFI/RFP asks the security, governance, and lock-in questions sales decks skip.

See the template — RFI $299 / RFP $699 →

Is this your product?

Claiming is free and lets you correct your listing's data. Upgrade to Enhanced ($199/yr) for a vendor-maintained mark, a richer profile, and lead routing.