Token Kiosk
Models

Model notes

Per-model quirks to be aware of.

These are documented behaviors that differ from the generic OpenAI request shape.

kimi-k2.6

kimi-k2.6 ignores temperature, top_p, and penalty parameters.

MiniMax models

MiniMax models ignore presence_penalty and frequency_penalty parameters.

gpt-oss-20b

gpt-oss-20b requires max_tokens ≥ 500 to produce output. Accepts reasoning_effort.

gpt-5.x

gpt-5.5 and gpt-5.6-sol are reasoning models. The gateway translates max_tokens to the model's max_completion_tokens field for you, but temperature, top_p, penalty params, and logprobs are silently dropped — only default sampling is supported.

thinking and reasoning_effort are model-specific

These are not gateway-wide flags, and the gateway does not reject them with an error. Support is per-model:

  • claude-* accepts thinking; reasoning content comes back as a reasoning field (see Anthropic (Claude) models).
  • gpt-oss-* and gpt-5.* accept reasoning_effort.
  • gpt-5.* silently drops thinking if sent.
  • Every other provider forwards these fields as-is to its upstream API, which may ignore or error on them depending on that provider's own request validation.

Omit both fields unless you're targeting one of the reasoning models above.

On this page