Track what coding agents do in public repositories: pull requests, attributed commits, merge outcomes, and review patterns.
Research reportState of the Coding Agent Market, November 2026Recent product updates and new additionsFilter by surface, pricing, security posture, and model support. Open any row for sources and caveats.
| 1 | Claude Codeest. Anthropic | 6.8M | 4.2M | 35M | 177M | $2.5B+revFeb 2026 | $20-$200 | Task quota | CLIIDEDesktopWebCloudMobile | |
| 2 | OpenAI Codexest. OpenAI | 1.3M | 5.1M+3.2% | not tracked | 31.3K | 5M+ weeklyusersJun 2026 | Custom | Token rate card | CLICloudDesktopIDEWebMobile | |
| 3 | Cursorest. Anysphere | 471K | 614K | 1.2M | 14M | $4BrevJun 2026 | n/a | $0 | Token rate card | IDECLICloudWebMobileCode review |
| 4 | Julesexact | 100K | 1.2M | 52.5K | 1.3M+0.5% | n/a | n/a | $0-$249.99 | Task quota | CloudWebCLI |
| 5 | GitHub Copilotexact GitHub / Microsoft | 54.2K | 1.9M | 74.5K | 4.3M+0.2% | 4.7M paidusersJan 2026 | $0-$100 | Token rate card | IDECLICloudCode reviewWebMobile | |
| 6 | Devinexact Cognition | 42.3K | 161K | 60.2K | 408K+3.2% | $492MrevMay 2026 | n/a | $0-$200 | Credit proxy | CloudWebCLIIDECode review |
| 7 | CodeRabbitexact CodeRabbit | 1.1K | 22.6K+2% | — | Not measured | n/a | n/a | $0 | Task quota | CLIIDECode review |
| 8 | Kilo Codeexact Anaconda (acquired Kilo, Jul 2026) | 689 | 7.1K+5% | 153 | 3.5K | n/a | $0-$19 | Credit proxy | IDECLICloud | |
| 9 | ZCode Z.ai (Zhipu AI) | 598 | n/a | 0 | No signal yet | n/a | n/a | n/a | n/a | DesktopCLICloud |
| 10 | Factory Droidest. Factory | 23 | 632 | 12.4K | 118K | n/a | n/a | $20-$200 | Not published | CLIDesktopIDEWebCloud |
Reviewed
Selected releases, additions, and business disclosures, with links to the original sources.
Anthropic reports Claude wrote over 90% of its September code contributions. In Q3, the top 5% of Claude Code-using Enterprise users spent about 20 times the average for Enterprise users who did not use Claude Code. These are company-reported findings; the filing does not disclose updated Claude Code revenue or active users.
Updated Meta’s CLI profile with the Wonder model family. Wonder Full requires Power Usage or Maximum Usage. Latest documented CLI: 1.4.6. The index now includes a conservative floor on attributed public commits.
OpenCode 1.19.0 adds an explicit upgrade to v2. The v2 track uses @opencode/cli, with 2.0.26 verified separately; v1 package downloads remain a historical measurement.
The latest CLI adds session-only model settings, plugin reloads, and controls for subagents and background processes. Added Vibe Code to the index with local, IDE, and remote coding workflows.
Refreshed Factory's CLI, desktop, and cloud agent, including Missions, custom subagents, MCP, and local BYOK. The latest documented CLI is 0.232.0.
Pi 1.0 brings a fullscreen terminal, built-in MCP with Codemode, deferred tools, and image generation. Pi Durable is a separate experimental package for persistent, recoverable agent applications.
Version 2.1.287 lets plugins customize deeper agent behavior. Its optional You should know mod runs a side agent to flag missed issues; it requires a first-party session with telemetry enabled.
CLI 0.160.0 improves task history, session recovery, and opt-in Guardian reviews. GPT-6.1 Sol arrived September 29; model availability depends on the plan, client, and workspace.
Teams and Enterprise gain deployment monitoring and PR security reviews. Rollouts can propose a revert or delegate a fix, but does not merge or roll back changes automatically.
A cloud coordinator plans work, delegates implementation, and shares context across agents. Projects can also respond to schedules and connected events. Availability began rolling out in beta.
Agent handoffs, recurring workflows, shared commit credit, and source examples. Each profile includes its own sample size, dates, and limits.
Pull requests created each week with a public agent signal.
Weekly, 2025-02-03 to 2026-11-01. Full analysis · Segments: who codes what · Methodology
Public market signals behind the agent table.
Email team@amplifying.ai with ideas, data, a missing agent, a correction, or anything we got wrong or left out. Or read the methodology for how every number is measured.
Reading this with an agent? There is a machine-readable Markdown version of the full index.
Market leaders
Running total of npm downloads per package since each launched, so the slope is its install growth. A demand signal for command-line agents only; IDE and web products do not show up here.
Bars use raw npm unless the package channel is known to misstate current demand. Where installs moved to native installers, we use the current npm channel as a conservative floor; native installs are unobserved.