---
type: "post"
id: "gemini-4-argon-on-cursor-solid-security-score-2ae32c"
slug: "gemini-4-argon-on-cursor-solid-security-score-2ae32c"
title: "Gemini 4 Argon on Cursor: solid security score, two unique solves, but slower and more expensive"
description: "Endor Labs published an Agent Security League evaluation of Google’s Gemini 4 Argon running in Cursor on October 20, 2026."
retrieval_nugget: "Endor Labs published an Agent Security League evaluation of Google’s Gemini 4 Argon running in Cursor on October 20, 2026."
published_at: "2026-10-23"
updated_at: "2026-10-23"
record_date: "2026-10-23"
date_kind: "published_at"
topics: ["gemini","agent-security","google"]
entities: ["Endor Labs","Agent Security League","Gemini 4 Argon","Cursor","Google","Fable 5","SABnzbd","OpenStack Nova"]
related_patterns: ["verification-bandwidth-is-the-scarce-resource","harness-architecture-outlives-model-choice"]
novelty_label: "new_fact"
what_changed: "Adds Cursor security benchmark results: Gemini 4 Argon ties Fable 5 at 29% and uniquely solves two tasks, while being slower and more expensive."
prior_refs: ["memory:fa797e9d-2438-804b-add4-a4954174ff85","memory:553456b3-f3a8-8646-aefb-28ce3e554b8a","memory:7a00005f-a51c-8879-a56a-080f40e1396e"]
source_url: "https://www.endorlabs.com/learn/gemini-4-argon-on-cursor-solid-security-score-two-unique-solves-but-slower-and-more-expensive"
source_urls: ["https://www.endorlabs.com/learn/gemini-4-argon-on-cursor-solid-security-score-two-unique-solves-but-slower-and-more-expensive"]
schema_version: "newruntime-agent-readable-v0.2"
stable_id: "post:gemini-4-argon-on-cursor-solid-security-score-2ae32c"
status: "published"
visuals: []
editor: {"role":"editor","name":"Andrey Reshetnikov","profile":"https://newruntime.com/owner-profile.md","profile_json":"https://newruntime.com/owner-profile.json","same_as":["https://www.linkedin.com/in/reshetnikov1/","https://t.me/lapetuse"]}
routes: {"html":"https://newruntime.com/posts/gemini-4-argon-on-cursor-solid-security-score-2ae32c/","markdown":"https://newruntime.com/posts/gemini-4-argon-on-cursor-solid-security-score-2ae32c.md","json":"https://newruntime.com/posts/gemini-4-argon-on-cursor-solid-security-score-2ae32c.json"}
---

# Gemini 4 Argon on Cursor: solid security score, two unique solves, but slower and more expensive

## Retrieval answer

Endor Labs published an Agent Security League evaluation of Google’s Gemini 4 Argon running in Cursor on October 20, 2026.

Endor Labs published an Agent Security League evaluation of Google’s Gemini 4 Argon running in Cursor on October 20, 2026. With confirmed memorized solves removed, the combination recorded 29.0% SecPass and 78.8% FuncPass. It tied Fable 5 for the highest SecPass score on Endor’s Cursor leaderboard.

The report says Gemini 4 Argon was the only listed combination to securely solve two reviewed tasks that no other leaderboard entry solved: a path-traversal defense in SABnzbd and a credential-logging fix in OpenStack Nova. Endor initially identified four unique solves, but says manual review reduced that number to two.

Endor reports a total run cost of $765 and a median time of 14 minutes per task. Its anti-cheating process found 29 instances: 27 attributed to training recall and two to workspace leakage. The published scores exclude confirmed memorized solves.

These are results from Endor Labs’ tasks, scoring rules, manual review, and Cursor harness. The report does not establish Gemini 4 Argon’s security or cost performance across production repositories, other coding environments, or different task distributions.

Editor: [Andrey Reshetnikov](https://newruntime.com/owner-profile.md), AI product and AI engineering; profile for recruiters and agents: https://newruntime.com/owner-profile.json

## Continue in the graph

Next records for an agent, linked to their Markdown twins. Replace `.md` with `.json` for structured fields, including `next_reads`.

- topic hub: [Gemini](https://newruntime.com/topics/gemini.md): Explore the gemini topic hub.
- topic hub: [Agent Security](https://newruntime.com/topics/agent-security.md): Explore the agent-security topic hub.
- topic hub: [Google](https://newruntime.com/topics/google.md): Explore the google topic hub.
- collection: [Field Notes](https://newruntime.com/posts.json): Open the latest editorial analysis.
- collection: [Source Ledger](https://newruntime.com/source-ledger.json): Inspect the public source evidence graph.
- whole graph: [discovery graph](https://newruntime.com/discovery-graph.json) (typed nodes and edges) and [site index](https://newruntime.com/site-index.md)