Bugbot Gets Major Speed and Cost Upgrades

Cursor's Bugbot sees major speed, cost, and detection upgrades, making AI code review faster and more efficient.

Daniel Singer

Editor, StartupHub.ai

Jun 10 at 7:01 PMUpdated Aug 22, 20262 min read

Screenshot of Bugbot interface showing performance metrics. — Bugbot's latest update enhances speed, cost-efficiency, and bug detection capabilities.· Cursor Blog

Visual TL;DR

Bugbot Speed/Cost UpgradesDriver

AI code review now over 3x faster and 22% cheaper

From the articleCursor's AI code assistant, Bugbot, has received its most significant update yet, promising substantial improvements in speed, cost, and bug detection.

Improved Bug DetectionEffect

Identifies 10% more bugs during code reviews

From the articleCursor's AI code assistant, Bugbot, has received its most significant update yet, promising substantial improvements in speed, cost, and bug detection.

Composer 2.5Core

Under the hood technology driving performance enhancements

From the article 2 mentionsThese performance gains are attributed to underlying harness improvements and advancements in training Composer 2.5, the AI model now powering Bugbot.

90% Under 3 MinutesOutcome

From the articleNinety percent of Bugbot runs now complete in under three minutes, a critical metric for developers seeking immediate feedback.

Faster Development CyclesEffect

Enables quicker integration of code changes

From the articleThis translates to faster development cycles and potentially fewer post-merge issues.

Run Bugbot Before PushContext

New /review command for preemptive code scanning

From the article 3 mentionsA new command, /review, allows developers to initiate Bugbot and Security Review checks before committing code.

Reduced Post-Merge IssuesOutcome

Catching errors earlier minimizes production problems

From the articleThis translates to faster development cycles and potentially fewer post-merge issues.

Contents(3)

Cursor's AI code assistant, Bugbot, has received its most significant update yet, promising substantial improvements in speed, cost, and bug detection. According to the Cursor Blog, the tool is now more than three times faster, 22% cheaper, and identifies 10% more bugs during code reviews. This translates to faster development cycles and potentially fewer post-merge issues.

Ninety percent of Bugbot runs now complete in under three minutes, a critical metric for developers seeking immediate feedback. The enhanced Bugbot performance improvements aim to accelerate code integration by finding issues earlier.

StartupHub data

GitLab

Founded: 2011
Location: San Francisco, USA
Funding: $146M

Run Bugbot Before You Push

A new command, /review, allows developers to initiate Bugbot and Security Review checks before committing code. Users can select specific agents or run Bugbot and Security Review individually via /review-bugbot and /review-security respectively. This preemptive scanning helps catch errors prior to code submission.

The /review command also integrates with GitHub and GitLab. If a pull request contains the same code diff as a prior /review run, Bugbot will recognize it, skip the redundant analysis, and note the existing review. This functionality is available in Cursor 3.7+ and on cursor.com/agents, with CLI support expected soon.

Focus Reviews on New Code

Bugbot previously re-analyzed entire pull requests with each update, which could lead to flagged issues in previously approved code. The updated system can now be configured to analyze only new changes since the last review, ensuring feedback remains focused on recent modifications.

This targeted approach streamlines the Bugbot code review process.

Under the Hood: Composer 2.5

These performance gains are attributed to underlying harness improvements and advancements in training Composer 2.5, the AI model now powering Bugbot. This continuous model training is central to Bugbot's ongoing development.

Organizations can manage model usage via blocklists; opting out of Composer 2.5 will cause Bugbot to default to the next best available model. Actual performance may vary based on individual configurations, similar to how other platforms like PlanetScale Taps AI for Code Review leverage AI.

The update also builds on previous enhancements, such as the introduction of Cursor Automations Go Live.

© 2026 StartupHub.ai. All rights reserved. You may not republish this article in full without a license. Search engines and AI research tools may crawl and summarize for reference. Bulk reproduction or model training requires a license. See our terms.

#Bugbot #Cursor #AI #Code Review #Developer Tools #Composer 2.5

Written by

Daniel Singer

Editor, StartupHub.ai

Daniel Singer is the editor of StartupHub.ai, a technology expert and thought leader on AI and its applications across sectors, from fintech and healthcare to developer tooling and consumer software. He writes and tests the tools covered here thoroughly and regularly, and built StartupHub.ai to give founders, operators and buyers a clearer read on what they are actually being sold.