Is the API OpenAI-compatible?

Yes. Set the base URL to https://api.pentestllm.ai/v1 in any OpenAI SDK or client, use your key as the API key, and send a security build id or auto/security as the model.

Does it work with the Anthropic SDK?

Yes. The Messages endpoint at https://api.pentestllm.ai/v1/messages takes Anthropic-format requests with the key in the x-api-key header or as a bearer token.

Per token, at the price of the build that answered, from a prepaid balance. Server tools add their own fee per search, fetch or container-hour. There is no subscription and no router fee.

What is the cheapest security model on the API?

On 2026-11-11 the lowest-priced security build costs $0.40 per million input tokens. The pricing page shows the current figure for every build.

Can I cap what the router spends?

Yes. Send "auto": {"tier_max": "balanced"} or "fast" in the request body and the router will not pick a build from a higher tier.

Do the models support tool calling?

Many do. Each model page lists the request parameters the build supports, including tools and response_format.

A cybersecurity LLM API

PentestLLM is a cybersecurity LLM API: 9 security-tuned builds behind OpenAI-compatible and Anthropic-compatible endpoints, billed per token from a prepaid balance. You create a key, set the base URL to https://api.pentestllm.ai/v1, and send the model id auto/security or the id of a single build.

Get a key Prices per token

shell

curl https://api.pentestllm.ai/v1/messages \
  -H "x-api-key: sk-ptl-live-…" \
  -H "anthropic-version: 2023-06-01" \
  -H "content-type: application/json" \
  -d '{
    "model": "auto/security",
    "max_tokens": 1024,
    "messages": [{"role": "user", "content": "Draft a detection rule for this log pattern and explain each condition."}]
  }'

Rows of server cabinets in a data centre

Endpoints and authentication

Chat completions: POST https://api.pentestllm.ai/v1/chat/completions, the OpenAI chat format. Most agents and SDKs use this one.
Responses: POST https://api.pentestllm.ai/v1/responses, the OpenAI Responses format.
Messages: POST https://api.pentestllm.ai/v1/messages, the Anthropic format, for the Anthropic SDK and Claude Code.
Key: A bearer token in the Authorization header, or the same key in x-api-key for Anthropic-format clients.

Every security build answers on all three routes at the same price. The docs have the full reference, including errors and streaming.

One model id that routes

The id auto/security is a router over the security builds. It classifies each request from what it carries, without an extra model call: a reasoning request, or tools together with code, goes to the frontier tier; tools, code blocks, images, a JSON schema or a prompt over 16,000 tokens go to the balanced tier; everything else goes to the fast tier. The response’s model field names the build that answered, and you pay that build’s price. There is no router fee.

Two controls matter for cost and consistency. An auto object with tier_max keeps a request out of the higher tiers. An x-session-id header keeps one build answering a session for 30 minutes after its last request, so an agent run does not change model halfway.

shell

curl https://api.pentestllm.ai/v1/chat/completions \
  -H "Authorization: Bearer sk-ptl-live-…" \
  -H "Content-Type: application/json" \
  -H "x-session-id: engagement-42" \
  -d '{
    "model": "auto/security",
    "auto": {"tier_max": "balanced"},
    "messages": [{"role": "user", "content": "Summarize these findings for the weekly status note."}]
  }'

Tools that run on the API side

Web search: $0.01 per search. The model searches and reads results inside one request.
Web fetch: $0.001 per page fetched. The model retrieves a page you or it names, such as a vendor advisory.
Code sandbox: $0.05 per container-hour, 1-minute minimum. Bash, JavaScript and Python in an isolated container.
Remote MCP: No tool fee. The API calls a remote MCP server you name in the request.

Server tools are declared in the request and billed with the tokens. They need a build that takes tool calls; each model page says whether it does.

Pricing, as an API should state it

Each build has an input price and an output price per million tokens, read from the live catalog. The table below is that catalog on 2026-11-11. The balance is prepaid from $10 by card or crypto; a request that would overdraw it is refused, not billed. The same figures are on the pricing page, and the models page adds context windows and supported parameters.

Who the API is for

It is for lawful, authorized security work: penetration testers, red teams, bug bounty hunters, CTF players, analysts and the people who build tools for them. You must own the systems you test or hold written permission. The acceptable use policy is the full statement, and the agent guides cover the common clients.

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

Yes. Set the base URL to https://api.pentestllm.ai/v1 in any OpenAI SDK or client, use your key as the API key, and send a security build id or auto/security as the model.
Yes. The Messages endpoint at https://api.pentestllm.ai/v1/messages takes Anthropic-format requests with the key in the x-api-key header or as a bearer token.
Per token, at the price of the build that answered, from a prepaid balance. Server tools add their own fee per search, fetch or container-hour. There is no subscription and no router fee.
On 2026-11-11 the lowest-priced security build costs $0.40 per million input tokens. The pricing page shows the current figure for every build.
Yes. Send "auto": {"tier_max": "balanced"} or "fast" in the request body and the router will not pick a build from a higher tier.
Many do. Each model page lists the request parameters the build supports, including tools and response_format.

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