Skip to content
Agent categories
Coding AI Agent
Don’t see your agent?

Compare how coding AI agents operate

Compare documented skills, published claims, and operating models.

Cursor (Anysphere) logo
Cursor (Anysphere)

4.6(310 G2 reviews)

AI coding editor (IDE) plus cloud coding agents that autonomously write, test, review and ship code across terminal, GitHub and Slack, for software engineering teams. Developed by Anysphere.

G2 reviewsBuyer-reported scores from G2 reviews. These are separate from G2 task evaluation results.

Ease of use
92
Quality of support
85

G2 evaluation not yet available

Request this product for a future evaluation.

Skills

Delegate Coding Tasks to Autonomous AgentsParallel Cloud Agents That Build, Test and Demo FeaturesAutomated Pr Review with Bugbot

Performance claims

Vendor-reported

In head-to-head evaluations, 93% of engineers select Cursor as their preferred AI coding tool.

Product strategy

Self-serveyes
Agent-callable (A2A)yes
Replit (Agent) logo
Replit (Agent)

4.5(397 G2 reviews)

Replit Agent turns natural-language prompts into full-stack web and mobile apps, handling coding, testing, database, auth and deployment, for non-developers, PMs and teams building software.

G2 reviewsBuyer-reported scores from G2 reviews. These are separate from G2 task evaluation results.

Ease of use
89
Quality of support
81

G2 evaluation not yet available

Request this product for a future evaluation.

Skills

Build Production-Ready Web and Mobile Apps from A Chat PromptRapid Prototyping of Internal Business AppsBuild Agents and Automate Workflows

Performance claims

Vendor-reported

3x faster and 10x more cost effective than Computer Use Models

Product strategy

Forward-deployed engineeringno
Self-serveyes
Agent-callable (A2A)no
Cognition (Devin) logo
Cognition (Devin)

4.5(29 G2 reviews)

Devin is an autonomous AI software engineer that plans, writes, tests and ships code inside a team's codebase, for enterprise engineering teams. Cognition acquired Windsurf (2025).

G2 reviewsBuyer-reported scores from G2 reviews. These are separate from G2 task evaluation results.

Ease of use
88
Quality of support
85

G2 evaluation not yet available

Request this product for a future evaluation.

Skills

Pr Review and Visual QaCode Migration and RefactoringIssue Triage and Bug Fixing

Performance claims

Vendor-reported

8x engineering time efficiency gain

Product strategy

Forward-deployed engineeringyes
Self-serveyes
Agent-callable (A2A)yes