Parameters by Upstream
Some request fields do not mean the same thing everywhere. This page states what the gateway does with each one on the way to each provider - passed through, translated into the provider's own equivalent, or dropped.
This is not a list of every accepted field; each endpoint page documents its own request body. Listed here are the fields where supported is not a yes or no, because the answer changes with the provider your model runs on.
reasoning_effort
Full guide, with values and cost impact →
| You call | Upstream | Behaviour | Beschreibung |
|---|---|---|---|
/v1/chat/completions | OpenAI-compatible (GPT, GLM, DeepSeek, Qwen, Kimi…) | unverändert | Forwarded to the provider as sent, with one exception: when a model's catalog entry lists the effort levels it accepts, a value below the lowest is raised to it and one above the highest steps down (GPT-6.1 Sol cannot turn reasoning off, so none and minimal run at low). Otherwise the gateway neither validates the value nor guarantees the provider honours it. Commonly minimal, low, medium and high, but the accepted set is the provider's: OpenAI also takes none on most models, GLM takes up to max. Check the provider's own reference for its range. |
/v1/chat/completions | Google Vertex - Gemini 4 | abgebildet | Becomes Gemini's thinkingLevel, kept inside the three levels Gemini 4 Argon accepts: none and minimal run at low (the model cannot turn thinking off), low/medium/high map straight across, xhigh and max run at high. An explicit google.thinking_config in the request wins over this mapping. |
/v1/chat/completions | Google Vertex - Gemini 3.x | abgebildet | Becomes Gemini's thinkingLevel: none and minimal both map to minimal, low/medium/high map straight across. An explicit google.thinking_config in the request wins over this mapping. |
/v1/chat/completions | Google Vertex - Gemini 2.5 | verworfen | Gemini 2.5 only understands thinkingBudget and rejects thinkingLevel outright, so the gateway sends neither and leaves the model at its default thinking behaviour. Use google.thinking_config with thinking_budget to control 2.5. |
/v1/chat/completions | Anthropic (Claude) | abgebildet | Wird zu Anthropics thinking, in der Form, die das Modell annimmt. Bei Claude-Modellen, die nur adaptives Denken annehmen (Claude Opus 4.7 und neuer, Sonnet 5 und neuer, Haiku 5.5), werden low/medium/high/xhigh/max zu thinking:{"type":"adaptive"} mit output_config.effort auf derselben Stufe, und minimal läuft als low. none schaltet das Denken in der eigenen Form des Modells ab: {"type":"disabled"} bei den meisten, {"type":"between_tools"} bei Sonnet 5.5, und adaptiv auf low bei Opus 5.5, das das Denken nicht abschalten kann. Andere Werte werden nicht weitergegeben. Bei älteren Claude-Modellen wird daraus stattdessen ein Budget: none und minimal schalten das Denken ab ({"type":"disabled"}); low/medium/high/xhigh/max werden zu einem Budget von 2,048 / 8,192 / 16,384 / 32,768 Tokens, die Umkehrung der Stufenzuordnung in der anderen Richtung. Ein abgeleitetes Budget wird unter max_tokens gehalten und entfällt ganz, wenn max_tokens Anthropics Minimum von 1,024 Tokens nicht fasst. Eine Effort-Angabe macht also nie aus einer funktionierenden Anfrage einen 400-Fehler beim Upstream. |
/v1/messages | OpenAI-compatible (GPT, GLM, DeepSeek, Qwen, Kimi…) | abgebildet | Derived from thinking.budget_tokens by bucket, when extended thinking is enabled and the budget is above zero: up to 2,048 tokens becomes low, up to 8,192 becomes medium, above that becomes high. A zero or absent budget sends nothing. For a model whose catalog entry lists the effort levels it accepts (GPT-6.1 Sol), the request's intent is followed instead: thinking disabled or between_tools runs at the lowest level (low), even when output_config.effort is also set; otherwise output_config.effort is sent as the effort. |
/v1/responses | OpenAI-compatible (GPT, GLM, DeepSeek, Qwen, Kimi…) | abgebildet | Read out of the Responses-API reasoning.effort object and sent as the flat reasoning_effort field. |
thinking / thinking_budget / enable_thinking
| You call | Upstream | Behaviour | Beschreibung |
|---|---|---|---|
/v1/chat/completions | Anthropic (Claude) | abgebildet | Das sind die Namen anderer Anbieter für denselben Regler, und das Gateway vereinheitlicht sie. thinking:{"type":"disabled"}, enable_thinking:false und thinking_budget:0 schalten das Denken ab, genauso wie reasoning_effort none bei diesem Modell. Bei älteren Claude-Modellen geben thinking_budget:N und thinking:{"type":"enabled","budget_tokens":N} das Budget unverändert weiter; ein Wert außerhalb des Bereichs wird vom Upstream abgelehnt statt still angepasst. Bei Modellen, die nur adaptives Denken annehmen, wird nie ein Budget gesendet: Es wird zu output_config.effort (bis 2,048 low, bis 8,192 medium, unter 32,768 high, ab 32,768 xhigh; nie max). thinking:{"type":"adaptive"} allein sendet keine Denk-Einstellung, und das Modell läuft mit seinem Standardverhalten. |
/v1/chat/completions | Google Vertex - Gemini 4 | abgebildet | Normalised into thinkingConfig.thinkingLevel, because Gemini 4 Argon rejects thinkingBudget: turning thinking off runs at low, and an explicit budget is folded into a level (up to 2,048 tokens low, up to 8,192 medium, above that high). An explicit google.thinking_config in the request still wins. |
/v1/chat/completions | Google Vertex - Gemini 3.x | abgebildet | Normalised into thinkingConfig: turning thinking off becomes thinkingLevel:"minimal", and an explicit budget becomes thinkingBudget. An explicit google.thinking_config in the request still wins over all of it. |
/v1/chat/completions | Google Vertex - Gemini 2.5 | abgebildet | Normalised into thinkingConfig.thinkingBudget - turning thinking off sends a budget of 0, and an explicit budget passes through. thinkingLevel is not used here because 2.5 rejects it. An explicit google.thinking_config in the request still wins. |
Every row is transcribed from the converter that implements it and pinned by a Go test, so this table cannot drift from the running gateway without a build failing.