Cybersecurity model API.
Call a security-tuned model from your own code or agent through an OpenAI-compatible API, and pay for it with prepaid credit.
Available today
Qwen Cyber 27B AWQ, DeepHat V1 7B, Orca Qwen 3.8 27B, Orca Nex N2.5 Mini, Hermes 4 14B, Hermes 4.3 36B, Foundation-Sec 1.1 8B. Each card below shows its price. The other 17 models below are not available yet.
How to start
- Buy credit. Sign in with a passkey (no email needed), open Account, then Model API, and add US$20, US$50 or US$100.
- Create a key. Name it and give it a spending limit. The key starts with
bo_api_and is shown once, so save it in your environment asBOUNTY_OPERATOR_API_KEY: the call below reads it from there. - Call it. Send a chat request to
https://bountyoperator.com/v1/chat/completionswith your key and the model ID, from your own code or any tool that uses the OpenAI chat API.
After an idle spell the first request waits up to 10 minutes while the model starts. You are not charged for the wait.
curl https://bountyoperator.com/v1/chat/completions \
-H "Authorization: Bearer $BOUNTY_OPERATOR_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "cyber-qwen-27b-awq",
"messages": [
{ "role": "user", "content": "Trace every external call and state update in withdraw()." }
],
"max_tokens": 2048,
"chat_template_kwargs": { "enable_thinking": false }
}'
- Works with tools that use the OpenAI chat API
- Prepaid credit, no subscription
- A spending limit on every key
- We don’t store your prompts or answers
Requests also pass to the selected hosting provider, whose data policies apply. Read the provider and data handling details.
All models and prices
USD per million tokens. Every request costs at least the model’s minimum. Switch the price view to estimate a review or an agent session.
Checking model availability…
When a model has been idle, its first request waits up to 10 minutes while a GPU starts and loads it. Requests after that answer in seconds while it stays warm. You are not charged for the wait. What to expect
Use Infron’s models with your own Infron key
Infron’s own USD rates per million tokens, billed by Infron. Separate from the prepaid catalog above.
Already have an Infron API key? All seven models in Infron’s cybersecurity stack are wired directly into Bounty Operator. Open any model in the review workbench with one click, or pass --provider infron in the CLI and MCP server.
GLM-5.3 Flash Cybersecurity
z-ai/glm-5.3-flash-cybersecurity
- Input
- $0.75
- Output
- $12.50
DeepSeek V4 Flash Vision Exp Cybersecurity
deepseek/deepseek-v4-flash-vision-exp-cybersecurity
- Input
- $0.75
- Output
- $2.50
Qwen3.8 27B Cybersecurity
qwen/qwen3.8-27b-cybersecurity
- Input
- $1.25–2.00
- Output
- $8.00–37.50
Gemma 4 26B A4B IT Cybersecurity
google/gemma-4-26b-a4b-it-cybersecurity
- Input
- $0.60
- Output
- $1.90
GLM-5.3 Flash Low-Refusal
z-ai/glm-5.3-flash-abliterated
- Input
- $1.25
- Output
- $4.00
NVIDIA Nemotron 3.5 Content Safety
nvidia/nemotron-3.5-content-safety
- Input
- $0.05
- Output
- $0.15
Infron rates checked 14 October 2026. Rates vary by the upstream provider used; ranges show the listed routes. Platform fees are separate. Check Infron’s current model rates and fees before using your own key.
One balance.
Every model.
Add US$20, US$50 or US$100.
No subscription required. Auto top-ups are off by default.
Pay for what you use
Each request costs the token price or the model’s minimum charge, whichever is higher. Requests with no confirmed usage are not charged.
Credit does not expire. The US$10 monthly Operator plan also includes US$2 in API credit per paid month.
For authorized security work. Models can make mistakes; check their answers. Model licenses and use
Ready to connect your agent or run a review?
Buy credit, create a key and call the Model API from your own code or agent. Or start a free review on your own AI provider API key.
Our application does not store prompts or responses. The hosting provider’s data policies also apply.