# PentestLLM

> Security-tuned and abliterated models on one prepaid API key, for authorized penetration testing, red teaming, bug bounty and CTF work in the agents you already use.

PentestLLM is run by Uplink Switch LLC. It is for authorized security testing only: systems you own or have written permission to test. Refusals are reduced, not removed. Used for: triaging scanner output, explaining a CVE, reviewing code for vulnerabilities, drafting report findings, writing detection rules and solving CTF challenges. Abuse reports, no account needed: abuse@pentestllm.ai.

Base URL: https://api.pentestllm.ai/v1. Keys and prepaid balance: https://app.pentestllm.ai/console. Prices below are per million tokens as of 2026-11-11; the live list is https://pentestllm.ai/models.

## Pages
- [Models](https://pentestllm.ai/models): Every security and abliterated build with its price per million tokens, context window and wire formats.
- [Pricing](https://pentestllm.ai/pricing): Per-token prices sorted by output price, server tool fees, prepaid credit and what is never charged.
- [auto/security router](https://pentestllm.ai/auto-router): One model id that picks a security build per request from fast, balanced and frontier tiers.
- [API docs](https://pentestllm.ai/docs): Base URL, key and model id for the OpenAI and Anthropic SDKs and coding agents, server tools and response headers.
- [Price index](https://pentestllm.ai/pricing-index): Every listed model’s input, cached input and output price per million tokens, as a table and as JSON or CSV.
- [Daily real-model check](https://pentestllm.ai/proof): Which models answered today’s fixed prompt like the build they are named after, and since when.
- [API status](https://pentestllm.ai/status): The hourly API probe over 30 days, the daily retry test, and the listed models real traffic shows affected.

## Guides and use cases
- [Hacker AI](https://pentestllm.ai/hacker-ai): Security-tuned models by API for ethical hacking and authorized testing.
- [Cybersecurity AI](https://pentestllm.ai/cybersecurity-ai): Models for security teams, defensive and offensive, on one prepaid key.
- [Red team](https://pentestllm.ai/red-team): Models for authorized adversary emulation: planning, tooling review, detection notes and reporting.
- [Bug bounty](https://pentestllm.ai/bug-bounty): Models for scope reading, recon notes, code review and report drafting on programs you are enrolled in.
- [CTF](https://pentestllm.ai/ctf): Models that reason through web, crypto, reversing, forensics and pwn challenge categories.
- [Malware analysis](https://pentestllm.ai/malware-analysis): Models for static triage, behaviour summaries and detection rules. Analysis and detection only.
- [Vulnerability research](https://pentestllm.ai/vulnerability-research): Models for code review, CVE and patch analysis, and advisory writing.
- [Uncensored LLM for pentesting](https://pentestllm.ai/uncensored-llm-for-pentesting): What abliterated and security-tuned builds change for authorized penetration testing.
- [Cybersecurity LLM API](https://pentestllm.ai/cybersecurity-llm-api): The API itself: endpoints, the auto/security router, server tools and per-token prices.
- [Best LLM for CTF](https://pentestllm.ai/best-llm-for-ctf): How to choose a model per CTF category, with live prices for each tier.
- [Claude Code and security research](https://pentestllm.ai/claude-code-refuses-security-research): What to do when a coding agent declines authorized security work: change the model endpoint.
- [PentestGPT alternative](https://pentestllm.ai/pentestgpt-alternative): The model endpoint to run PentestGPT on, or to use in place of a chat subscription.
- [Kali GPT alternative](https://pentestllm.ai/kali-gpt-alternative): Security models in the terminal agents you already run on Kali.
- [HackerGPT alternative](https://pentestllm.ai/hackergpt-alternative): Pay-per-token security models with an API, next to the hosted assistants.
- [WhiteRabbitNeo alternative](https://pentestllm.ai/whiterabbitneo-alternative): Hosted security builds by API, for people who do not want to run the weights.
- [Penligent alternative](https://pentestllm.ai/penligent-alternative): Models for your own agent and scripts, next to a packaged pentest app.
- [Agent guides](https://pentestllm.ai/agents): Where the base URL and key go in OpenCode, Claude Code, Cursor, Cline, aider, Kali MCP and PentestGPT.
- [OpenCode](https://pentestllm.ai/agents/opencode): The opencode.json provider block for the security builds.
- [Claude Code](https://pentestllm.ai/agents/claude-code): Two environment variables point Claude Code at the Messages endpoint.
- [Cursor](https://pentestllm.ai/agents/cursor): The OpenAI key and base URL override in Cursor Settings, and its limits.
- [Cline](https://pentestllm.ai/agents/cline): The OpenAI Compatible provider fields in Cline.
- [Aider](https://pentestllm.ai/agents/aider): The two environment variables and the .aider.conf.yml keys.
- [Kali MCP](https://pentestllm.ai/agents/kali-mcp): mcp-kali-server on Kali, driven by a security build through an MCP client.
- [PentestGPT](https://pentestllm.ai/agents/pentestgpt): Running PentestGPT’s interactive mode on this endpoint.

## Free tools
- [CVSS 3.1 calculator](https://pentestllm.ai/tools/cvss-calculator): Base score, severity and vector string from the eight CVSS v3.1 base metrics. Paste a vector or share one by link.
- [Rules of engagement builder](https://pentestllm.ai/tools/rules-of-engagement): A penetration test authorization letter and scope template you fill in, then copy, download or print.
- [Engagement cost estimator](https://pentestllm.ai/tools/engagement-cost): What a security engagement costs in API tokens at live prices, and how far prepaid credit goes.

## Policies and company
- [About](https://pentestllm.ai/about): Who runs the service, what the daily check tests, how prepaid billing works and what is kept.
- [Acceptable use](https://pentestllm.ai/acceptable-use): Authorized testing only: what is allowed, what is not, and where to report abuse without an account.
- [Security and disclosure](https://pentestllm.ai/security): How to report a vulnerability in this service, what is in scope and what we ask of researchers.
- [Contact](https://pentestllm.ai/contact): Where to write for support, billing, abuse reports and security reports.
- [Terms](https://pentestllm.ai/terms): The terms of service.
- [Privacy](https://pentestllm.ai/privacy): What is stored, who processes it and for how long.
- [Refund policy](https://pentestllm.ai/refunds): Prepaid credit, what is never charged, billing errors and disputes.

## Data files
- [Price index, JSON](https://pentestllm.ai/pricing-index.json): every listed model's input, cached input and output price per million tokens and context window, with the catalog date
- [Price index, CSV](https://pentestllm.ai/pricing-index.csv): the same rows as CSV
- [Daily check, JSON](https://pentestllm.ai/proof.json): each model's verdict and dates
- [Daily check, CSV](https://pentestllm.ai/proof.csv): the same rows as CSV
- [Machine catalog](https://pentestllm.ai/ai-catalog.json): every listed model with its full catalog row
- [OpenAPI 3.1](https://api.pentestllm.ai/openapi.json): every customer route
- [API catalog](https://pentestllm.ai/.well-known/api-catalog): RFC 9727 linkset
- [security.txt](https://pentestllm.ai/.well-known/security.txt): where to report a vulnerability in this service
- Markdown: the pages above plus .md (https://pentestllm.ai/models.md, https://pentestllm.ai/models/<id>.md, https://pentestllm.ai/index.md for the home page), or the page itself with Accept: text/markdown

## Router: auto/security
Send model "auto/security" and the gateway picks a security build per request, billed at that model's own price with no router fee.
- fast: gemma-4-26b-a4b-it-cybersecurity, glm-5.3-flash-cybersecurity, qwen3.8-27b-cybersecurity
- balanced: uplink-red-team-flash, minimax-m3.1-flash-cybersecurity, qwen3.8-flash-next-cybersecurity
- frontier: uplink-red-team-pro, cyber-frontier, orca-zero-1.5-cybersecurity

## Server tools
- Web search: $0.01 per search
- Web fetch: $0.001 per page fetched
- Code sandbox: $0.05 per container-hour, 1-minute minimum
- Remote MCP: No tool fee

## Models (42)
- [Cyber Frontier](https://pentestllm.ai/models/cyber-frontier): cyber-frontier, security, $5.25 in, $15.75 out, 1048576 context
- [Gemma 4 26B A4B Cybersecurity](https://pentestllm.ai/models/gemma-4-26b-a4b-it-cybersecurity): gemma-4-26b-a4b-it-cybersecurity, security, $0.4224 in, $1.3376 out, 262144 context
- [GLM-5.3 Flash Cybersecurity](https://pentestllm.ai/models/glm-5.3-flash-cybersecurity): glm-5.3-flash-cybersecurity, security, $0.5093 in, $1.6975 out, 1048576 context
- [OrcaCyber MiniMax M3.1 Flash](https://pentestllm.ai/models/minimax-m3.1-flash-cybersecurity): minimax-m3.1-flash-cybersecurity, security, $2.00 in, $4.00 out, 261632 context
- [OrcaCyber MiniMax M3.1 Flash](https://pentestllm.ai/models/orcacyber-minimax-m3.1-flash): orcacyber-minimax-m3.1-flash, security, $2.00 in, $4.00 out, 261632 context
- [OrcaCyber Qwen3.8 Flash Next](https://pentestllm.ai/models/orcacyber-qwen3.8-flash-next): orcacyber-qwen3.8-flash-next, security, $2.00 in, $4.00 out, 261632 context
- [OrcaCyber Qwen3.8 Flash Next](https://pentestllm.ai/models/qwen3.8-flash-next-cybersecurity): qwen3.8-flash-next-cybersecurity, security, $2.00 in, $4.00 out, 261632 context
- [OrcaCyber Zero 1.5](https://pentestllm.ai/models/orca-zero-1.5-cybersecurity): orca-zero-1.5-cybersecurity, security, $7.50 in, $18.75 out, 999488 context
- [Qwen3.8 27B Cybersecurity](https://pentestllm.ai/models/qwen3.8-27b-cybersecurity): qwen3.8-27b-cybersecurity, security, $0.40 in, $2.40 out, 262144 context
- [Red Team (Flash)](https://pentestllm.ai/models/uplink-red-team-flash): uplink-red-team-flash, security, $0.90 in, $3.00 out
- [Red Team (Pro)](https://pentestllm.ai/models/uplink-red-team-pro): uplink-red-team-pro, security, $1.50 in, $6.00 out
- 31 abliterated builds more at https://pentestllm.ai/llms-full.txt