Is hacker AI legal to use?

PentestLLM is sold for lawful, authorized security work. Testing a system without permission is illegal in most places whatever tool you use, so keep your testing to systems you own or have written permission to test.

Will the models refuse hacking questions?

Far less often than a general assistant. The builds are tuned or abliterated for security work, so they answer questions about vulnerabilities, tooling and detection that other models decline. That is reduced refusals, not zero.

Which model should I start with?

Send the model id auto/security. The router reads each request and picks a fast, balanced or frontier security build, and the response names the build that answered. You can pin a single build later from the models page.

How much does hacker AI cost here?

You pay per token from a prepaid balance. On 2026-11-11 the lowest-priced security build costs $0.40 per million input tokens, and the smallest top-up is $10 by card or crypto. The pricing page lists every build.

Can I use it inside Claude Code or OpenCode?

Yes. The API takes OpenAI-format and Anthropic-format requests, so OpenCode, Cline, Cursor, aider and Claude Code can all point at it. The agent guides give the exact settings for each tool.

Hacker AI for authorized testing

Hacker AI is a language model that helps with ethical hacking: it reads scanner output, explains vulnerabilities, reviews code and drafts findings for systems you are allowed to test. PentestLLM sells API access to 9 security-tuned builds on one prepaid key from $10, so the model runs inside the terminal and coding agents you already use.

Get a key Prices per token

shell

curl https://api.pentestllm.ai/v1/chat/completions \
  -H "Authorization: Bearer sk-ptl-live-…" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "auto/security",
    "messages": [{"role": "user", "content": "Rank these scanner findings by severity and say what to verify first."}]
  }'

Rows of server cabinets in a data centre

What people mean by hacker AI

The phrase covers three different things: a chat assistant with a security persona, an autonomous agent that runs tools for you, and the language model underneath both. PentestLLM is the third. It is an API, so you bring the interface, whether that is OpenCode, Claude Code, Cursor, a Python script or plain curl.

The reason a separate model matters is refusals. A general assistant is trained to decline a wide band of security questions, and that band includes things a tester asks every working day: what a finding means, how to confirm it on a host in scope, what a proof of concept in a public advisory is doing. Security-tuned and abliterated builds are trained or adjusted to answer those questions. That is reduced refusals, not a promise that a model answers everything.

What the models do on an engagement

Triage: Turn raw scanner and tool output into a ranked list, with what to verify first and what is likely noise.
Explain: Walk through a CVE advisory: what is affected, how to confirm exposure on a host you are testing, and what the patch changes.
Review: Read source code or configuration for vulnerability classes and point at the lines that matter.
Map: Tie findings to OWASP, CWE, MITRE ATT&CK or the control framework your client reports against.
Report: Draft a finding with impact, evidence placeholders and remediation steps in the structure of your template.
Detect: Write a Sigma or YARA rule for behaviour you observed, so the defending team can catch it next time.

How it differs from a chat subscription

A subscription gives you a chat window and a monthly cap. An API key gives you the model wherever your work happens. Requests go to https://api.pentestllm.ai/v1 in the OpenAI format or to the Messages endpoint in the Anthropic format, and you pay for the tokens each request used from a prepaid balance. There is no seat licence and no renewal.

One model id, auto/security, routes each request to a fast, balanced or frontier build depending on what it carries, and bills it at that build’s own price. The API can also run web search, page fetch, a code sandbox and remote MCP servers on its side, so an agent without those tools still gets them. The agent guides show where the base URL and key go in each tool.

Hacking AI, AI hacker, hacker AI: same need

People search for hacking AI, AI hacker and hacker AI and mean the same thing: an assistant that will engage with offensive security instead of changing the subject. What separates the options is where the model runs and how you pay. If you want a hosted chat, the HackerGPT alternative page compares the hosted assistants. If you want the model in your own tools, start with the models and the price list.

Authorized use, in plain terms

The service is for lawful security work on systems you own or have written permission to test: client engagements under a signed scope, bug bounty programs you are enrolled in, CTF events, labs and your own infrastructure. A model that answers security questions does not make an unauthorized test legal. The acceptable use policy lists what is not allowed, and says a key may be suspended until proof of authorization is shown.

9 security builds, priced per token

Read from the live catalog on 2026-11-11. Prices are US dollars per million tokens, paid from a prepaid balance that starts at $10. The id auto/security has no price of its own: each request is billed at the build that answered.

Build	Input	Output	Context
Qwen3.8 27B Cybersecurity	$0.40	$2.40	262,144
Gemma 4 26B A4B Cybersecurity	$0.4224	$1.3376	262,144
GLM-5.3 Flash Cybersecurity	$0.5093	$1.6975	1,048,576
Red Team (Flash)	$0.90	$3.00	not published
Red Team (Pro)	$1.50	$6.00	not published
OrcaCyber MiniMax M3.1 Flash	$2.00	$4.00	261,632
OrcaCyber Qwen3.8 Flash Next	$2.00	$4.00	261,632
Cyber Frontier	$5.25	$15.75	1,048,576
OrcaCyber Zero 1.5	$7.50	$18.75	999,488

Full price list and tool fees · Every model · API docs

Questions

PentestLLM is sold for lawful, authorized security work. Testing a system without permission is illegal in most places whatever tool you use, so keep your testing to systems you own or have written permission to test.
Far less often than a general assistant. The builds are tuned or abliterated for security work, so they answer questions about vulnerabilities, tooling and detection that other models decline. That is reduced refusals, not zero.
Send the model id auto/security. The router reads each request and picks a fast, balanced or frontier security build, and the response names the build that answered. You can pin a single build later from the models page.
You pay per token from a prepaid balance. On 2026-11-11 the lowest-priced security build costs $0.40 per million input tokens, and the smallest top-up is $10 by card or crypto. The pricing page lists every build.
Yes. The API takes OpenAI-format and Anthropic-format requests, so OpenCode, Cline, Cursor, aider and Claude Code can all point at it. The agent guides give the exact settings for each tool.

One key, from $10

PentestLLM is for systems you own or have written permission to test. The acceptable use policy says what is not allowed.

Get a key

usage, 1,000 agent turns on Qwen3.8 27B Cybersecurity

Input, 8,000 tokens at $0.40 per million: $0.0032
Output, 1,000 tokens at $2.40 per million: $0.0024
One turn: $0.0056
1,000 turns: $5.60

output price per million tokens, 9 builds

$1.3376 Gemma 4 26B A4B CybersecurityOrcaCyber Zero 1.5 $18.75