Skip to main content
POST

Authorizations

Authorization
string
header
required

Bearer authentication header of the form Bearer <token>, where <token> is your auth token.

Body

application/json
model
enum<string>
required

AnyFast model ID.

Available options:
qwen3.6-plus
Example:

"qwen3.6-plus"

messages
object[]
required

Conversation messages. Content may be text or an array of text, image, and video parts.

Minimum array length: 1
max_completion_tokens
integer
default:65536

Maximum number of generated tokens. Qwen3.6-Plus supports up to 64K output tokens.

Required range: 1 <= x <= 65536
temperature
number
default:1

Sampling temperature. Valid range is [0, 2). Set only one of temperature and top_p.

Required range: 0 <= x < 2
Example:

1

top_p
number

Nucleus sampling probability. Valid range is (0, 1]. Set only one of temperature and top_p.

Required range: 0 < x <= 1
stream
boolean
default:false

If true, stream partial message deltas using Server-Sent Events.

stream_options
object

Options for streaming responses. Only relevant when stream is true.

stop
string[]

An array of stop sequences. The string form is not supported by the current AnyFast endpoint.

seed
integer

Seed for more repeatable sampling.

Required range: 0 <= x <= 2147483647
presence_penalty
number
default:0

Penalizes new tokens based on whether they appear in the text so far.

Required range: -2 <= x <= 2
frequency_penalty
number
default:0

Penalizes repeated tokens based on their frequency in the text so far.

Required range: -2 <= x <= 2
enable_thinking
boolean

Controls reasoning mode. With the OpenAI SDK, pass this provider field through extra_body. Use non-thinking mode for workflows that require structured output or tool calling when required by the upstream model.

tools
object[]

Application-defined functions or supported tools the model may call.

tool_choice

Controls whether the model calls tools or selects a function.

Available options:
none,
auto,
required
response_format
object

Structured output configuration. Use non-thinking mode when required by the upstream model.

Response

Completion generated successfully. The response is a complete completion when stream is false, or a completion chunk when stream is true.

id
string
required
Example:

"chatcmpl-abc123"

object
enum<string>
required
Available options:
chat.completion
created
integer
required

Unix timestamp.

model
string
required
Example:

"qwen3.6-plus"

choices
object[]
required
usage
object