# Feedback Bench: Coding Agents, Ranked by their users’ feedback

Built 2026-10-01. Window 2026-08-31 to 2026-09-27. Web page: https://feedbackbench.com/. Built by Enterpret (https://www.enterpret.com), which runs the same feedback analysis on a company's own customer feedback.

Feedback Bench ranks 17 coding agents from what their users say in public: Reddit, X, G2 and Trustpilot. Every post is labelled against 63 criteria in 9 areas. The Feedback Score is 100 × √(Popularity × Customer love): Popularity is how many people post about the agent (log share of authors, leader = 1); Customer love is how those posts rate it against the category (0.5 = category norm). Full method: [https://feedbackbench.com/method.md](https://feedbackbench.com/method.md).

## Ranking

Rank shows “=N” when rank ranges overlap. Criteria better / worse than peers counts the 63 criteria where the agent's 95% interval lies wholly above / below 0.5.

| Rank | Agent | Maker | Feedback Score | 95% interval | Rank range | Popularity | Customer love | Customer love 95% interval | Criteria better / worse than peers |
|---|---|---|---|---|---|---|---|---|---|
| 1 | [Claude Code](https://feedbackbench.com/agents/claude-code.md) | Anthropic | 70.4 | 69.9–70.9 | 1–1 | 0.986 | 0.503 | 0.496–0.510 | 14 / 10 |
| 2 | [OpenAI Codex](https://feedbackbench.com/agents/codex.md) | OpenAI | 67.4 | 66.9–67.8 | 2–2 | 1.000 | 0.454 | 0.448–0.460 | 7 / 21 |
| 3 | [OpenCode](https://feedbackbench.com/agents/opencode.md) | Anomaly (open source) | 66.0 | 65.4–66.7 | 3–3 | 0.773 | 0.564 | 0.553–0.575 | 7 / 7 |
| 4 | [Cursor](https://feedbackbench.com/agents/cursor.md) | Anysphere | 59.7 | 59.0–60.5 | 4–4 | 0.708 | 0.504 | 0.491–0.516 | 5 / 10 |
| =5 | [Devin](https://feedbackbench.com/agents/devin.md) | Cognition | 55.9 | 55.1–56.5 | 5–6 | 0.514 | 0.607 | 0.591–0.622 | 9 / 2 |
| =5 | [Google Antigravity](https://feedbackbench.com/agents/antigravity.md) | Google | 54.9 | 54.0–55.7 | 5–6 | 0.650 | 0.463 | 0.449–0.477 | 3 / 18 |
| 7 | [Pi](https://feedbackbench.com/agents/pi.md) | Earendil Works (open source) | 52.7 | 51.9–53.3 | 7–7 | 0.456 | 0.608 | 0.591–0.623 | 14 / 0 |
| =8 | [GitHub Copilot](https://feedbackbench.com/agents/copilot.md) | GitHub | 43.0 | 42.2–43.7 | 8–10 | 0.348 | 0.532 | 0.513–0.551 | 1 / 1 |
| =8 | [Cline](https://feedbackbench.com/agents/cline.md) | Cline (open source) | 42.9 | 42.2–43.6 | 8–10 | 0.333 | 0.553 | 0.535–0.570 | 3 / 0 |
| =8 | [Zed](https://feedbackbench.com/agents/zed.md) | Zed Industries | 42.4 | 41.7–43.0 | 8–10 | 0.362 | 0.496 | 0.481–0.511 | 2 / 2 |
| =11 | [Factory](https://feedbackbench.com/agents/factory.md) | Factory | 35.6 | 34.9–36.2 | 11–12 | 0.232 | 0.545 | 0.525–0.563 | 2 / 0 |
| =11 | [Amp](https://feedbackbench.com/agents/amp.md) | Amp | 35.3 | 34.7–35.8 | 11–12 | 0.222 | 0.561 | 0.544–0.577 | 5 / 0 |
| 13 | [Kiro](https://feedbackbench.com/agents/kiro.md) | AWS | 29.5 | 29.0–30.0 | 13–13 | 0.185 | 0.470 | 0.453–0.485 | 0 / 2 |
| 14 | [Conductor](https://feedbackbench.com/agents/conductor.md) | Melty Labs | 26.4 | 26.1–26.7 | 14–14 | 0.136 | 0.510 | 0.499–0.521 | 0 / 0 |
| 15 | [Warp](https://feedbackbench.com/agents/warp.md) | Warp | 23.2 | 23.0–23.5 | 15–15 | 0.107 | 0.506 | 0.495–0.516 | 0 / 0 |
| 16 | [Grok Build](https://feedbackbench.com/agents/grok-build.md) | xAI | 12.6 | 12.4–12.7 | 16–16 | 0.031 | 0.517 | 0.506–0.527 | 0 / 0 |
| 17 | [Augment Code](https://feedbackbench.com/agents/augment.md) | Augment | 9.7 | 9.7–9.7 | 17–17 | 0.019 | 0.495 | 0.491–0.499 | 0 / 0 |

## Top quadrant

Rule: popularity of at least 0.5 and customer love of at least 0.5. At the edge: the customer love 95% interval still includes 0.5.

| Agent | Rank | Score | Popularity | Customer love | Customer love 95% interval | At the edge |
|---|---|---|---|---|---|---|
| Claude Code | 1 | 70.4 | 0.99 | 0.503 | 0.496–0.510 | yes |
| OpenCode | 3 | 66.0 | 0.77 | 0.564 | 0.553–0.575 | no |
| Cursor | 4 | 59.7 | 0.71 | 0.504 | 0.491–0.516 | yes |
| Devin | =5 | 55.9 | 0.51 | 0.607 | 0.591–0.622 | no |

## Areas

Reading per agent and area: Better than peers, Typical, Worse than peers, or Too few posts (under 30 rated author-weeks).

| Agent | [Paying and limits](https://feedbackbench.com/criteria/paying.md) | [Setting up and connecting](https://feedbackbench.com/criteria/setup.md) | [Choosing models](https://feedbackbench.com/criteria/models.md) | [Instructing and context](https://feedbackbench.com/criteria/context.md) | [Doing the work](https://feedbackbench.com/criteria/work.md) | [Checking and finishing](https://feedbackbench.com/criteria/checking.md) | [Interface and sessions](https://feedbackbench.com/criteria/interface.md) | [Reliability and speed](https://feedbackbench.com/criteria/reliability.md) | [Account and support](https://feedbackbench.com/criteria/account.md) |
|---|---|---|---|---|---|---|---|---|---|
| Claude Code | Worse than peers | Better than peers | Better than peers | Typical | Typical | Typical | Better than peers | Better than peers | Worse than peers |
| OpenAI Codex | Worse than peers | Worse than peers | Worse than peers | Typical | Typical | Typical | Worse than peers | Worse than peers | Better than peers |
| OpenCode | Better than peers | Typical | Better than peers | Typical | Typical | Typical | Typical | Better than peers | Typical |
| Cursor | Typical | Typical | Typical | Better than peers | Better than peers | Better than peers | Typical | Worse than peers | Worse than peers |
| Devin | Better than peers | Typical | Better than peers | Typical | Better than peers | Typical | Better than peers | Typical | Too few posts |
| Google Antigravity | Typical | Worse than peers | Typical | Worse than peers | Worse than peers | Worse than peers | Worse than peers | Better than peers | Typical |
| Pi | Better than peers | Better than peers | Better than peers | Typical | Better than peers | Too few posts | Better than peers | Better than peers | Typical |
| GitHub Copilot | Better than peers | Typical | Better than peers | Typical | Typical | Typical | Typical | Typical | Typical |
| Cline | Better than peers | Typical | Better than peers | Typical | Typical | Too few posts | Better than peers | Better than peers | Typical |
| Zed | Typical | Worse than peers | Too few posts | Too few posts | Worse than peers | Typical | Worse than peers | Better than peers | Typical |
| Factory | Better than peers | Typical | Better than peers | Too few posts | Better than peers | Too few posts | Typical | Too few posts | Too few posts |
| Amp | Better than peers | Typical | Better than peers | Typical | Better than peers | Too few posts | Better than peers | Typical | Better than peers |
| Kiro | Worse than peers | Too few posts | Worse than peers | Too few posts | Typical | Too few posts | Too few posts | Too few posts | Typical |
| Conductor | Too few posts | Too few posts | Too few posts | Too few posts | Typical | Too few posts | Typical | Too few posts | Too few posts |
| Warp | Too few posts | Too few posts | Too few posts | Too few posts | Too few posts | Too few posts | Typical | Too few posts | Too few posts |
| Grok Build | Too few posts | Too few posts | Too few posts | Too few posts | Too few posts | Too few posts | Too few posts | Too few posts | Too few posts |
| Augment Code | Too few posts | Too few posts | Too few posts | Too few posts | Too few posts | Too few posts | Too few posts | Too few posts | Too few posts |

## Pages and data

- Agents: [Claude Code](https://feedbackbench.com/agents/claude-code.md), [OpenAI Codex](https://feedbackbench.com/agents/codex.md), [OpenCode](https://feedbackbench.com/agents/opencode.md), [Cursor](https://feedbackbench.com/agents/cursor.md), [Devin](https://feedbackbench.com/agents/devin.md), [Google Antigravity](https://feedbackbench.com/agents/antigravity.md), [Pi](https://feedbackbench.com/agents/pi.md), [GitHub Copilot](https://feedbackbench.com/agents/copilot.md), [Cline](https://feedbackbench.com/agents/cline.md), [Zed](https://feedbackbench.com/agents/zed.md), [Factory](https://feedbackbench.com/agents/factory.md), [Amp](https://feedbackbench.com/agents/amp.md), [Kiro](https://feedbackbench.com/agents/kiro.md), [Conductor](https://feedbackbench.com/agents/conductor.md), [Warp](https://feedbackbench.com/agents/warp.md), [Grok Build](https://feedbackbench.com/agents/grok-build.md), [Augment Code](https://feedbackbench.com/agents/augment.md)
- Criteria: one page per area and criterion under https://feedbackbench.com/criteria/ (list in [https://feedbackbench.com/llms.txt](https://feedbackbench.com/llms.txt))
- All data as JSON: [https://feedbackbench.com/data/coding.json](https://feedbackbench.com/data/coding.json)
- Ranking as CSV: [https://feedbackbench.com/data/rankings.csv](https://feedbackbench.com/data/rankings.csv)
- Every agent × criterion as CSV: [https://feedbackbench.com/data/criteria.csv](https://feedbackbench.com/data/criteria.csv)
- Top requests per agent and criterion as CSV: [https://feedbackbench.com/data/requests.csv](https://feedbackbench.com/data/requests.csv)