GPT-5.6 Cyber API Pricing
GPT-5.6 Cyber is OpenAI's model for authorised vulnerability research, exploit validation and security testing, sold only to defenders approved into its Daybreak programme. OpenAI still publishes the rate card: $12.50/M input, $1.25/M cached input, $15.63/M cache writes and $75.00/M output - three times GPT-5.6 Sol's input and nearly four times its output. Checked by hand against developers.openai.com.
Run the numbers.
The rate card is public even though access is not, so the arithmetic is real - it applies to an account OpenAI has approved for Daybreak. Cache misses bill at the cache-write rate, as on every GPT-5.6 model; Batch is not available for this model.
Real-world presets.
Codebase vulnerability sweep
Crash-to-exploit validation
Scanner finding triage
Pentest write-up
Counted with the o200k_base BPE encoding — the same tokenizer the API uses — entirely in your browser. The encoder loads on your first keystroke (~1 MB, one time); your text never leaves this page.
| Model | Input /M | Output /M | Effective blended | Context | Best for |
|---|---|---|---|---|---|
| GPT-5.6 Cyber Current | $12.50 cache $1.25 | $75.00 | $9.53 agentic 92/8 | 400K | Authorised vulnerability research |
| GPT-5.5 Cyber | $12.50 cache $1.25 | $75.00 | $9.01 −5% | not listed | The earlier Daybreak cyber model |
| GPT-5.6 Sol | $4.00 cache $0.40 | $20.00 | $2.73 −3.5× · promo | 1.05M | Daybreak Blue: general model with cyber safeguards |
| GPT-6 Astra | $10.00 cache $1.00 | $50.00 | $6.82 −28% | 1.05M | OpenAI's generally available flagship |
| Claude Mythos 5.1 | $10.00 cache $0.25 | $50.00 | $6.26 −34% | 1M | Restricted cyber-capable model from another lab |
Frequently asked.
Practical GPT-5.6 Cyber pricing questions - what the card says, and who can actually be billed on it.
Q · 01 What is GPT-5.6 Cyber priced at? +
gpt-5.6-cyber at $12.50/M input, $1.25/M cached input, $15.63/M cache writes and $75.00/M output. There is no long-context tier: the model accepts at most 272K input tokens.Q · 02 Can I sign up for it? +
Q · 03 What are Daybreak Blue and Daybreak Red? +
gpt-daybreak-blue-latest points to GPT-5.6 Sol, a general model with safeguards for defensive work; gpt-daybreak-red-latest points to this model. OpenAI says the aliases move as new Daybreak models ship, with pricing matching each underlying model - so a bill on an alias can change without your code changing.Q · 04 How does it compare with GPT-5.6 Sol on price? +
$4.00/M input and $20.00/M output (a promotional card OpenAI keeps at least through November 21, 2026). GPT-5.6 Cyber is about three times Sol's input and nearly four times its output. Under our standard blend the gap is $9.53/M against $2.73/M.