Is PentestLLM a replacement for Strix?

Not directly. Strix is an agent and PentestLLM is a model API. Strix needs a model provider, and this endpoint can be that provider, or you can use the same key in a different agent.

The open-source tool is free under the Apache 2.0 licence, as its README stated in November 2026. It needs a language model, which is billed by whichever provider you connect. The hosted platform listed Pro at $29 per seat per month with pentests billed separately.

Which models does Strix support?

Its docs in November 2026 say it uses LiteLLM and supports more than 100 providers, recommend GLM-5.3 through a gateway, and list models from OpenAI, Anthropic, Google, DeepSeek and Moonshot, plus local models.

Have you tested Strix on PentestLLM?

We have not run a full Strix scan on this endpoint. The settings on this page follow the Strix and LiteLLM documentation for OpenAI-compatible servers, and you should check the first run yourself.

What does PentestLLM cost?

Per token from a prepaid balance with a $10 minimum top-up by card and no subscription. On 2026-11-12 input prices for security builds start at $0.40 per million tokens.

A Strix alternative on the model side

Strix is an open-source AI penetration testing agent, and like every agent it needs a language model behind it. PentestLLM is the model side: 9 security-tuned builds on one prepaid key from $10, to run Strix on or to use in another agent.

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."}]
  }'

What Strix is, as of November 2026

Strix’s README on GitHub, read on 11 November 2026, describes an open-source AI penetration testing tool released under the Apache 2.0 licence. It runs from the command line and needs two things: Docker, and an API key for a language model provider. The README lists an HTTP interception proxy, browser and shell execution, reconnaissance, static and dynamic analysis, multi-agent orchestration, a headless mode for CI and support for MCP servers.

The company also sells a hosted platform. Its pricing page on the same day listed Pro at $29 per seat per month with pentests billed separately per test, and an Enterprise plan at a custom price that includes VPC or on-premises deployment and custom model support. The open-source tool is free. The model behind it is not.

Strix does not include a model. Its docs say it uses LiteLLM and supports more than 100 providers, with the model set in STRIX_LLM as provider/model-name, the key in LLM_API_KEY and an optional LLM_API_BASE. The docs recommend GLM-5.3 through a gateway and list models from OpenAI, Anthropic, Google, DeepSeek and Moonshot. They also note that the frontier OpenAI, Anthropic and Gemini models need the vendor’s cyber access program before they can be used for security testing.

Why the model matters for Strix

Refusals: An autonomous agent sends many offensive security requests in a row. A model that declines one can stall the run.
Access programs: Strix’s docs say the frontier models from three large vendors need a cyber access program. A security-tuned build needs a key.
Cost: Multi-agent runs use many tokens. Per-token billing on a build you choose keeps the cost of each run visible.
Context: Long runs carry long histories. The largest window listed here today is 1M tokens.

Pointing Strix at this endpoint

Strix’s docs say it works with any OpenAI-compatible server and give LLM_API_BASE as the setting, with Ollama as the worked example. They do not show a generic example, so the model string here follows LiteLLM’s documented convention for OpenAI-compatible endpoints: the prefix openai/ in front of the model id. It uses a named build, because the id has to be one this API serves.

We have not run a full Strix scan against this endpoint and do not claim every Strix feature works on it. Treat these three lines as the documented starting point and check the first run. The agents we have verified step by step are in the agent guides.

shell

export STRIX_LLM="openai/gemma-4-26b-a4b-it-cybersecurity"
export LLM_API_KEY="sk-ptl-live-…"
export LLM_API_BASE="https://api.pentestllm.ai/v1"

What PentestLLM serves

PentestLLM does not ship an agent. It serves 9 security-tuned builds and the abliterated and uncensored builds beside them, behind an OpenAI-compatible endpoint at https://api.pentestllm.ai/v1 and an Anthropic-compatible Messages endpoint. The builds are tuned or abliterated to answer security questions other models decline. That is reduced refusals, not zero, and it is for authorized testing only.

Billing is per token from a prepaid balance that starts at $10 by card, with no subscription. On 2026-11-12 input prices for security builds start at $0.40 per million tokens, and the table below is the live catalog.

Or use a different agent

If you do not need Strix’s scanner and want an assistant that plans, runs tools with your approval and keeps notes, a general coding agent on a security build does that. OpenCode takes a provider block in one config file, Claude Code takes two environment variables, and the Kali MCP guide connects tool execution on a Kali machine. PentestGPT is another open-source pentest agent with its own guide.

When Strix is the better choice

If you want an autonomous scanner that finds and validates vulnerabilities in your own application, with a proxy, a browser and a sandbox already wired together, use Strix; a model API gives you none of that. If you want it hosted, with pull request reviews and scheduled tests, its platform is the product to look at. And if your organization already has access to a frontier model through a cyber access program, Strix’s docs list those models directly.

PentestLLM is for the cases where the model is the limiting part. Strix’s README carries the same warning we do: use it only on systems you own or have explicit written permission to test. Our acceptable use policy has the detail.

9 security builds, priced per token

Read from the live catalog on 2026-11-12. 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

Not directly. Strix is an agent and PentestLLM is a model API. Strix needs a model provider, and this endpoint can be that provider, or you can use the same key in a different agent.
The open-source tool is free under the Apache 2.0 licence, as its README stated in November 2026. It needs a language model, which is billed by whichever provider you connect. The hosted platform listed Pro at $29 per seat per month with pentests billed separately.
Its docs in November 2026 say it uses LiteLLM and supports more than 100 providers, recommend GLM-5.3 through a gateway, and list models from OpenAI, Anthropic, Google, DeepSeek and Moonshot, plus local models.
We have not run a full Strix scan on this endpoint. The settings on this page follow the Strix and LiteLLM documentation for OpenAI-compatible servers, and you should check the first run yourself.
Per token from a prepaid balance with a $10 minimum top-up by card and no subscription. On 2026-11-12 input prices for security builds start at $0.40 per million tokens.

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