Models

The research API's models come in two teams. The blue team needs only a sign-in; the red team needs identity verification and acceptance of the acceptable-use policy. The default is qwen3.8-flash-next-whitehacker.

Blue team

General-purpose models. Signing in is enough; the free allowance applies.

Model	Availability	Source	Attribution check
`qwen/qwen3.8-flash`	—	Qwen	The upstream must return its own model id
`z-ai/glm-5.3-flash`	—	zai-org	The upstream must return its own model id

Conditions of use

Signed in authenticated-principal
Free allowance remaining available-free-quota
Guardrails on every request guardrails

Red team

Verification required

Unrestricted models for security research. Only researchers who verified their identity with a card and accepted the acceptable-use policy can use them.

Model	Availability	Source	Attribution check
`qwen3.8-flash-next-whitehacker` Default	—	Qwen3.8-Flash-Next-CYBERSECURITY-NVFP4	The upstream must return its own model id
`glm5.3-flash`	—	GLM	Verified by the shape and content of the response

Conditions of use

Signed in authenticated-principal
Identity verified with a card verified-ekyc-card
Acceptable-use policy (AUP) accepted aup-consent
AML/CTF screening current current-aml-ctf-clearance
Trust route current current-trust-route
Approved research scope approved-research-scope
Free allowance remaining available-free-quota
Guardrails on every request guardrails

Free allowance

Requests per day: 1000
Max output tokens: 32768
Max input characters: 524288

Shared by both teams. Paid credits are added under Billing.

How to call it

Shell

curl -s -X POST https://api.kotoba.cloud/v1/chat/completions \
  -H "Authorization: Bearer $KOTOBA_API_TOKEN" \
  -H "content-type: application/json" \
  -d '{"model":"qwen3.8-flash-next-whitehacker","messages":[{"role":"user","content":"…"}]}'

Pass the id from the table as `model`, unchanged. The quickstart walks from issuing a token to the first request.