Feature Request: Add Support for OpenAI/Azure Responses API

Aperta
#25 0 commenti 0 reazioni 0 assegnatari Vedi su GitHub

Nessuno ha ancora preso questa issue.

Valutazione

Difficoltà
5/5
Tempo stimato
Più di una settimana
Idoneità per principianti
35/100
Tipo di issue
Funzionalità
Chiarezza
Abbastanza chiara
Stato di attività
Tranquilla
Stack tecnologico
php
Ambito
api

Direzione di ricerca

Inizia individuando l’integrazione attuale di $client->chat()->create e la gestione dei modelli descritta nell’issue. Definisci come devono coesistere il percorso Chat Completions esistente e il nuovo percorso /responses, quindi verifica che GPT-5 Pro utilizzi Responses mentre i modelli esistenti rimangano compatibili.

Scritto dal modello di indicizzazione a partire dal testo della issue.

Descrizione

The current implementation only supports the Chat Completions API. This prevents the use of newer reasoning models such as gpt-5-pro, which are only available through the Responses API.

I would like to request support for the OpenAI/Azure Responses API, including automatic routing based on model capabilities.

⸻

Problem

When using Azure OpenAI with a GPT-5 Pro deployment, requests fail with the following error:

Invalid API request
The requested operation is unsupported.

Investigation shows that the library always sends requests to the Chat Completions endpoint:

POST /chat/completions

However, Azure GPT-5 Pro only supports the Responses API and does not support Chat Completions.

As a result, users cannot use GPT-5 Pro even though the deployment and credentials are configured correctly.

⸻

Current Behavior

The implementation assumes all chat-capable models support Chat Completions:

$client->chat()->create([
'model' => $model,
'messages' => $messages,
]);

This works for:

  • GPT-4o
  • GPT-4.1
  • GPT-5 Mini
  • GPT-5

but fails for:

  • GPT-5 Pro

⸻

Expected Behavior

The library should support both APIs:

Chat Completions

POST /chat/completions

Responses API

POST /responses

The framework should be able to route requests to the appropriate API depending on model capabilities.

Example:

if ($modelCapabilities->supportsResponsesOnly()) {
$client->responses()->create(...);
} else {
$client->chat()->create(...);
}

⸻

Proposed Solution

Option 1: Native Responses API Support

Introduce a dedicated Responses API client:

$client->responses()->create([
'model' => 'gpt-5-pro',
'input' => $messages,
]);

This gives users direct access to the newer API surface.

⸻

Option 2: Automatic Capability Routing

Add model capability detection and transparently choose the correct endpoint.

Example:

$client->generate([
'model' => 'gpt-5-pro',
'messages' => $messages,
]);

Internally:

if ($model->supportsResponses()) {
return $responsesClient->create(...);
}
return $chatClient->create(...);

This provides the best developer experience and future-proofs the implementation as more models migrate to Responses API.

⸻

Additional Considerations

The Responses API introduces features that are becoming the primary interface for modern reasoning models:

  • GPT-5 Pro support
  • Future reasoning model compatibility
  • Unified tool calling
  • Structured outputs
  • Multi-modal input/output
  • Reasoning controls

Supporting only Chat Completions will increasingly limit compatibility with newer OpenAI and Azure OpenAI models.

⸻

Environment

Provider

Azure OpenAI

API Version

2024-12-01-preview

Working Model

gpt-5-mini

Failing Model

gpt-5-pro

Error

Invalid API request
The requested operation is unsupported.

⸻

Benefit

Adding Responses API support would enable compatibility with GPT-5 Pro and future reasoning-focused models while maintaining backward compatibility with existing Chat Completions integrations.

Lingua principale
PHP
Stelle
49
Fork
11
Metriche di merge delle PR
Nessuna PR unita negli ultimi 30g

Preparare l'ambiente

Questo progetto non fornisce container di sviluppo, Dockerfile né guida per i contributori, quindi l'ambiente è a tuo carico: parti dal suo README e consulta la nostra guida al primo contributo per i passaggi generali.

Come iniziare

  1. Leggi tutta la issue e poi la guida ai contributi del progetto.
  2. Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
  3. Fai un fork del repository e lavora su un branch.
  4. Apri una pull request che faccia riferimento al numero della issue.

Altre issue di hyperf/odin

Tutte le issue di hyperf/odin

Issue simili

Altre issue su PHP

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.