Who is CyberSecEvalTest for?

CyberSecEvalTest is an open-source project built for security researchers.

Does CyberSecEvalTest have a free plan?

Yes — CyberSecEvalTest is free to use.

What platforms does CyberSecEvalTest run on?

CyberSecEvalTest runs on the web. It can also be self-hosted.

Is CyberSecEvalTest still maintained?

Unverified. CyberSecEvalTest has not been re-checked since it entered the index, so there is no finding either way — and only a positive finding would say otherwise.

What projects are similar to CyberSecEvalTest?

Similar projects tracked by PulseGate include constraintloop, trajectory-judge, and DeFake.

Who makes CyberSecEvalTest?

CyberSecEvalTest is developed by Meta.

How long has CyberSecEvalTest been around?

CyberSecEvalTest first shipped in 2023.

Back to the index

CyberSecEvalTest

Name: CyberSecEvalTest
Author: Meta

huggingface.coLLM eval & observability

CyberSecEval is a Hugging Face Space by Meta that tests large language models on various cybersecurity challenges. It runs standardized evaluations, displays results on a leaderboard, and provides visual analytics. The tool helps AI researchers and security professionals understand and mitigate risks associated with deploying LLMs in sensitive environments.

FreeWebSelf-hostedCloud-managed

Visit huggingface.co

Overview

4 features

CyberSecEvalTest sits in PulseGate's LLM eval & observability category. It focuses on assessing how vulnerable or capable large language models are at cybersecurity-related tasks. It is built as an open-source project for security researchers. CyberSecEvalTest is free to use. CyberSecEvalTest is available on the web, and it can be self-hosted.

Meta builds and maintains CyberSecEvalTest, and it first shipped in 2023. Among its 4 catalogued features are LLM Evaluation, Cybersecurity Tests, and leaderboard.

Summary written by a language model from the project’s public pages.

LLM Evaluation
Cybersecurity Tests
Leaderboard
Visual Analysis

Built with & integrations

Hosting

AWS

AI providers

meta_llamalocal_oss

Runs on

BrowserSelf-hostedCloud-managed

Detected from

AWS: x-amz-cf-id header · x-amz-cf-pop header · via header
local_oss: huggingface in the HTML

Trust & compliance

Verified signals

HTTPSFree tier

Indexing history

What PulseGate has recorded for this listing

Indexed22 Jul · 21:05 UTC
facebook/CyberSecEval verified against its public source
Source: PulseGate · Open

Frequently asked questions about CyberSecEvalTest

What does CyberSecEvalTest do?: CyberSecEvalTest focuses on assessing how vulnerable or capable large language models are at cybersecurity-related tasks. It is catalogued under LLM eval & observability on PulseGate.
Who is CyberSecEvalTest for?: CyberSecEvalTest is an open-source project built for security researchers.
Does CyberSecEvalTest have a free plan?: Yes — CyberSecEvalTest is free to use.
What platforms does CyberSecEvalTest run on?: CyberSecEvalTest runs on the web. It can also be self-hosted.
Is CyberSecEvalTest still maintained?: Unverified. CyberSecEvalTest has not been re-checked since it entered the index, so there is no finding either way — and only a positive finding would say otherwise.
What projects are similar to CyberSecEvalTest?: Similar projects tracked by PulseGate include constraintloop, trajectory-judge, and DeFake.constraintloop trajectory-judge DeFake
Who makes CyberSecEvalTest?: CyberSecEvalTest is developed by Meta.
How long has CyberSecEvalTest been around?: CyberSecEvalTest first shipped in 2023.

At a glance

Platforms: Web
Languages: English
Built for: security researchers
Model: Open source
Solves: Assessing how vulnerable or capable large language models are at cybersecurity-related tasks.

Registered as

Hugging Face Spaces: facebook/CyberSecEval

Developer

Meta

Live coverage

Identity confidence: Medium · 84.8
Indexed: 22 Jul 2026
Lifecycle: Alive
Last seen: 22 Jul 2026

Identity audit (13)

Slug: facebook-cyberseceval-huggingface-co
Lifecycle state recorded: 22 Jul 2026
Verification state: Indexed for public listing
Listing state: Listed: yes
Index status: Included in index
Latest evidence snapshot: 22 Jul 2026
Timeline basis: Indexed-at chronology. This listing's first-seen date was written by the catalog backfill, not observed here, so it is not treated as a sighting.
Name from: Derived from the project's own page and URL.
Category from: Assigned by a language model.
Summary from: Written by a language model from public pages.
Languages from: Detected by a language model from page content.
Last updated: 24 Jul 2026
Canonical URL: https://huggingface.co/spaces/facebook/CyberSecEval

Ship this? Send a correction — no account, and you get a link to follow it.

Also in LLM eval & observability

Same category — not a similarity match