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 | Descrizione |
|---|---|---|---|
/v1/chat/completions | OpenAI-compatible (GPT, GLM, DeepSeek, Qwen, Kimi…) | inoltrato | 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 | convertito | 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 | convertito | 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 | scartato | 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) | convertito | Diventa il thinking di Anthropic, nella forma che il modello accetta. Sui modelli Claude che accettano solo il ragionamento adattivo (Claude Opus 4.7 e successivi, Sonnet 5 e successivi, Haiku 5.5), low/medium/high/xhigh/max diventano thinking:{"type":"adaptive"} con output_config.effort allo stesso livello, e minimal gira come low. none disattiva il ragionamento nella forma propria del modello: {"type":"disabled"} sulla maggior parte, {"type":"between_tools"} su Sonnet 5.5, e adattivo a low su Opus 5.5, che non può disattivare il ragionamento. Qualsiasi altro valore non viene inoltrato. Sui modelli Claude precedenti diventa invece un budget: none e minimal disattivano il ragionamento ({"type":"disabled"}); low/medium/high/xhigh/max diventano un budget di 2,048 / 8,192 / 16,384 / 32,768 token, l'inverso della suddivisione usata nella direzione opposta. Un budget derivato resta sotto max_tokens e non viene inviato affatto quando max_tokens non contiene il minimo di 1,024 token di Anthropic, quindi chiedere un livello di effort non trasforma mai una richiesta funzionante in un errore 400 del fornitore. |
/v1/messages | OpenAI-compatible (GPT, GLM, DeepSeek, Qwen, Kimi…) | convertito | 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…) | convertito | 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 | Descrizione |
|---|---|---|---|
/v1/chat/completions | Anthropic (Claude) | convertito | Sono i nomi che altri fornitori danno allo stesso controllo, e il gateway li normalizza. thinking:{"type":"disabled"}, enable_thinking:false e thinking_budget:0 disattivano il ragionamento, come reasoning_effort none su quel modello. Sui modelli Claude precedenti, thinking_budget:N e thinking:{"type":"enabled","budget_tokens":N} inoltrano il budget invariato; un valore fuori intervallo viene rifiutato dal fornitore anziché corretto in silenzio. Sui modelli che accettano solo il ragionamento adattivo non viene mai inviato un budget: diventa output_config.effort (fino a 2,048 low, fino a 8,192 medium, sotto 32,768 high, da 32,768 xhigh; mai max). thinking:{"type":"adaptive"} da solo non invia alcuna impostazione di ragionamento, e il modello gira con il comportamento predefinito. |
/v1/chat/completions | Google Vertex - Gemini 4 | convertito | 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 | convertito | 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 | convertito | 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.