Directory

Every agent. Every verdict.

Sorted by Verdict Score. Click any agent for the full breakdown — sub-scores, failure modes, and what we tested.

1Verified11Provisional0Needs retest0Retired
All categoriesCoding AgentsBrowser AgentsBusiness Automation AgentsResearch AgentsSales AgentsCustomer Support AgentsData Analysis AgentsContent AgentsNo-Code Agent Builders
Provisional scores in effect

AgentVerdict is in public build mode. Initial agent scores are research-based provisional estimates, not the result of completed benchmark runs. They cannot be used as final proof and they are not eligible for "Verdict Certified" status. Verified scores will replace provisional scores as controlled test runs are published on the runs page.

Scores marked with * on the directory are provisional. Look for the Verified status badge for backed-by-evidence scores.

#1
84
/100
Claude Code
Verified
Anthropic's terminal-native coding agent. Reads, edits, and runs code across full repositories with explicit user approval for destructive actions.
Coding Agents
#2
82*
/100
Cursor
Provisional
AI-first IDE forked from VS Code with inline completions, chat, and an agent mode that executes multi-file edits.
Coding Agents
#3
78*
/100
Windsurf
Provisional
AI-first IDE from the team behind Codeium. Cascade agent flow blends inline edits with multi-step actions.
Coding Agents
#4
76*
/100
Aider
Provisional
Open-source command-line coding assistant that pairs with you in your terminal and commits changes via git.
Coding Agents
#5
75*
/100
OpenAI Codex CLI
Provisional
OpenAI's open-source terminal coding agent. Runs locally, executes code, and edits files with user approval.
Coding Agents
#6
74*
/100
Cline
Provisional
Open-source VS Code extension that runs an agentic coding loop with your choice of model.
Coding Agents
#7
73*
/100
GitHub Copilot Coding Agent
Provisional
GitHub's coding agent that picks up assigned issues and opens PRs from the GitHub UI, in addition to in-IDE Copilot features.
Coding Agents
#8
71*
/100
Replit Agent
Provisional
Replit's in-browser agent that scaffolds and ships apps end-to-end inside the Replit workspace.
Coding AgentsNo-Code Agent Builders
#9
70*
/100
Zapier Agents
Provisional
Zapier's AI agents layer sits on top of its 7,000+ app integrations, letting non-engineers build agents that act across SaaS tools.
Business Automation Agents
#10
68*
/100
OpenHands
Provisional
Open-source coding agent (formerly OpenDevin) from All Hands AI. Runs locally or self-hosted, model-agnostic.
Coding Agents
#11
67*
/100
Devin
Provisional
Cognition's autonomous AI software engineer. Runs in its own cloud sandbox and aims to complete tickets end-to-end.
Coding Agents
#12
64*
/100
Manus
Provisional
General-purpose AI agent that runs in a managed cloud sandbox to browse the web and complete multi-step tasks.
Browser AgentsResearch Agents