Add OpenRouterModel as OpenAIChatModel subclass

pydantic_ai_slim/pydantic_ai/providers/openrouter.py

@@ -81,13 +81,21 @@ def __init__(self, *, api_key: str) -> None: ...
     @overload
     def __init__(self, *, api_key: str, http_client: httpx.AsyncClient) -> None: ...
 
+    @overload
+    def __init__(self, *, api_key: str, http_referer: str, x_title: str) -> None: ...
+
+    @overload
+    def __init__(self, *, api_key: str, http_referer: str, x_title: str, http_client: httpx.AsyncClient) -> None: ...
+
     @overload
     def __init__(self, *, openai_client: AsyncOpenAI | None = None) -> None: ...
 
     def __init__(
         self,
         *,
         api_key: str | None = None,
+        http_referer: str | None = None,
+        x_title: str | None = None,
         openai_client: AsyncOpenAI | None = None,
         http_client: httpx.AsyncClient | None = None,
     ) -> None:
@@ -98,10 +106,20 @@ def __init__(
                 'to use the OpenRouter provider.'
             )
 
+        attribution_headers: dict[str, str] = {}
+        if http_referer := http_referer or os.getenv('OPENROUTER_HTTP_REFERER'):
+            attribution_headers['HTTP-Referer'] = http_referer
+        if x_title := x_title or os.getenv('OPENROUTER_X_TITLE'):
+            attribution_headers['X-Title'] = x_title
+
         if openai_client is not None:
             self._client = openai_client
         elif http_client is not None:
-            self._client = AsyncOpenAI(base_url=self.base_url, api_key=api_key, http_client=http_client)
+            self._client = AsyncOpenAI(
+                base_url=self.base_url, api_key=api_key, http_client=http_client, default_headers=attribution_headers
+            )
         else:
             http_client = cached_async_http_client(provider='openrouter')
-            self._client = AsyncOpenAI(base_url=self.base_url, api_key=api_key, http_client=http_client)
+            self._client = AsyncOpenAI(
+                base_url=self.base_url, api_key=api_key, http_client=http_client, default_headers=attribution_headers
+            )

tests/models/cassettes/test_openrouter/test_openrouter_errors_raised.yaml

@@ -0,0 +1,161 @@
+interactions:
+- request:
+    headers:
+      accept:
+      - application/json
+      accept-encoding:
+      - gzip, deflate
+      connection:
+      - keep-alive
+      content-length:
+      - '158'
+      content-type:
+      - application/json
+      host:
+      - openrouter.ai
+    method: POST
+    parsed_body:
+      messages:
+      - content: Be helpful.
+        role: system
+      - content: Tell me a joke.
+        role: user
+      model: google/gemini-2.0-flash-exp:free
+      stream: false
+    uri: https://openrouter.ai/api/v1/chat/completions
+  response:
+    headers:
+      access-control-allow-origin:
+      - '*'
+      connection:
+      - keep-alive
+      content-type:
+      - application/json
+      permissions-policy:
+      - payment=(self "https://checkout.stripe.com" "https://connect-js.stripe.com" "https://js.stripe.com" "https://*.js.stripe.com"
+        "https://hooks.stripe.com")
+      referrer-policy:
+      - no-referrer, strict-origin-when-cross-origin
+      transfer-encoding:
+      - chunked
+      vary:
+      - Accept-Encoding
+    parsed_body:
+      error:
+        code: 429
+        message: Provider returned error
+        metadata:
+          provider_name: Google
+          raw: 'google/gemini-2.0-flash-exp:free is temporarily rate-limited upstream. Please retry shortly, or add your own
+            key to accumulate your rate limits: https://openrouter.ai/settings/integrations'
+      user_id: user_2wT5ElBE4Es3R4QrNLpZiXICmQP
+    status:
+      code: 429
+      message: Too Many Requests
+- request:
+    headers:
+      accept:
+      - application/json
+      accept-encoding:
+      - gzip, deflate
+      connection:
+      - keep-alive
+      content-length:
+      - '158'
+      content-type:
+      - application/json
+      host:
+      - openrouter.ai
+    method: POST
+    parsed_body:
+      messages:
+      - content: Be helpful.
+        role: system
+      - content: Tell me a joke.
+        role: user
+      model: google/gemini-2.0-flash-exp:free
+      stream: false
+    uri: https://openrouter.ai/api/v1/chat/completions
+  response:
+    headers:
+      access-control-allow-origin:
+      - '*'
+      connection:
+      - keep-alive
+      content-type:
+      - application/json
+      permissions-policy:
+      - payment=(self "https://checkout.stripe.com" "https://connect-js.stripe.com" "https://js.stripe.com" "https://*.js.stripe.com"
+        "https://hooks.stripe.com")
+      referrer-policy:
+      - no-referrer, strict-origin-when-cross-origin
+      transfer-encoding:
+      - chunked
+      vary:
+      - Accept-Encoding
+    parsed_body:
+      error:
+        code: 429
+        message: Provider returned error
+        metadata:
+          provider_name: Google
+          raw: 'google/gemini-2.0-flash-exp:free is temporarily rate-limited upstream. Please retry shortly, or add your own
+            key to accumulate your rate limits: https://openrouter.ai/settings/integrations'
+      user_id: user_2wT5ElBE4Es3R4QrNLpZiXICmQP
+    status:
+      code: 429
+      message: Too Many Requests
+- request:
+    headers:
+      accept:
+      - application/json
+      accept-encoding:
+      - gzip, deflate
+      connection:
+      - keep-alive
+      content-length:
+      - '158'
+      content-type:
+      - application/json
+      host:
+      - openrouter.ai
+    method: POST
+    parsed_body:
+      messages:
+      - content: Be helpful.
+        role: system
+      - content: Tell me a joke.
+        role: user
+      model: google/gemini-2.0-flash-exp:free
+      stream: false
+    uri: https://openrouter.ai/api/v1/chat/completions
+  response:
+    headers:
+      access-control-allow-origin:
+      - '*'
+      connection:
+      - keep-alive
+      content-type:
+      - application/json
+      permissions-policy:
+      - payment=(self "https://checkout.stripe.com" "https://connect-js.stripe.com" "https://js.stripe.com" "https://*.js.stripe.com"
+        "https://hooks.stripe.com")
+      referrer-policy:
+      - no-referrer, strict-origin-when-cross-origin
+      transfer-encoding:
+      - chunked
+      vary:
+      - Accept-Encoding
+    parsed_body:
+      error:
+        code: 429
+        message: Provider returned error
+        metadata:
+          provider_name: Google
+          raw: 'google/gemini-2.0-flash-exp:free is temporarily rate-limited upstream. Please retry shortly, or add your own
+            key to accumulate your rate limits: https://openrouter.ai/settings/integrations'
+      user_id: user_2wT5ElBE4Es3R4QrNLpZiXICmQP
+    status:
+      code: 429
+      message: Too Many Requests
+version: 1

tests/models/cassettes/test_openrouter/test_openrouter_map_messages_reasoning.yaml

@@ -0,0 +1,96 @@
+interactions:
+- request:
+    headers:
+      accept:
+      - application/json
+      accept-encoding:
+      - gzip, deflate
+      connection:
+      - keep-alive
+      content-length:
+      - '133'
+      content-type:
+      - application/json
+      host:
+      - openrouter.ai
+    method: POST
+    parsed_body:
+      messages:
+      - content: Who are you. Think about it.
+        role: user
+      model: anthropic/claude-3.7-sonnet:thinking
+      stream: false
+    uri: https://openrouter.ai/api/v1/chat/completions
+  response:
+    headers:
+      access-control-allow-origin:
+      - '*'
+      connection:
+      - keep-alive
+      content-length:
+      - '4024'
+      content-type:
+      - application/json
+      permissions-policy:
+      - payment=(self "https://checkout.stripe.com" "https://connect-js.stripe.com" "https://js.stripe.com" "https://*.js.stripe.com"
+        "https://hooks.stripe.com")
+      referrer-policy:
+      - no-referrer, strict-origin-when-cross-origin
+      transfer-encoding:
+      - chunked
+      vary:
+      - Accept-Encoding
+    parsed_body:
+      choices:
+      - finish_reason: stop
+        index: 0
+        logprobs: null
+        message:
+          content: "I am Claude, an AI assistant created by Anthropic. I'm a large language model designed to be helpful,
+            harmless, and honest.\n\nI don't have consciousness or sentience like humans do - I'm a sophisticated text prediction
+            system trained on a large dataset of human text. I don't have personal experiences, emotions, or a physical existence.
+            \n\nMy purpose is to assist you with information, tasks, and conversation in a helpful way, while acknowledging
+            my limitations. I have knowledge cutoffs, can occasionally make mistakes, and don't have the ability to access
+            the internet or take actions in the physical world.\n\nIs there something specific you'd like to know about me
+            or how I can assist you?"
+          reasoning: |-
+            This question is asking me about my identity. Let me think about how to respond clearly and accurately.
+
+            I am Claude, an AI assistant created by Anthropic. I'm designed to be helpful, harmless, and honest in my interactions with humans. I don't have a physical form - I exist as a large language model running on computer hardware. I don't have consciousness, sentience, or feelings in the way humans do. I don't have personal experiences or a life outside of these conversations.
+
+            My capabilities include understanding and generating natural language text, reasoning about various topics, and attempting to be helpful to users in a wide range of contexts. I have been trained on a large corpus of text data, but my training data has a cutoff date, so I don't have knowledge of events that occurred after my training.
+
+            I have certain limitations - I don't have the ability to access the internet, run code, or interact with external systems unless given specific tools to do so. I don't have perfect knowledge and can make mistakes.
+
+            I'm designed to be conversational and to engage with users in a way that's helpful and informative, while respecting important ethical boundaries.
+          reasoning_details:
+          - format: anthropic-claude-v1
+            index: 0
+            signature: ErcBCkgICBACGAIiQHtMxpqcMhnwgGUmSDWGoOL9ZHTbDKjWnhbFm0xKzFl0NmXFjQQxjFj5mieRYY718fINsJMGjycTVYeiu69npakSDDrsnKYAD/fdcpI57xoMHlQBxI93RMa5CSUZIjAFVCMQF5GfLLQCibyPbb7LhZ4kLIFxw/nqsTwDDt6bx3yipUcq7G7eGts8MZ6LxOYqHTlIDx0tfHRIlkkcNCdB2sUeMqP8e7kuQqIHoD52GAI=
+            text: |-
+              This question is asking me about my identity. Let me think about how to respond clearly and accurately.
+
+              I am Claude, an AI assistant created by Anthropic. I'm designed to be helpful, harmless, and honest in my interactions with humans. I don't have a physical form - I exist as a large language model running on computer hardware. I don't have consciousness, sentience, or feelings in the way humans do. I don't have personal experiences or a life outside of these conversations.
+
+              My capabilities include understanding and generating natural language text, reasoning about various topics, and attempting to be helpful to users in a wide range of contexts. I have been trained on a large corpus of text data, but my training data has a cutoff date, so I don't have knowledge of events that occurred after my training.
+
+              I have certain limitations - I don't have the ability to access the internet, run code, or interact with external systems unless given specific tools to do so. I don't have perfect knowledge and can make mistakes.
+
+              I'm designed to be conversational and to engage with users in a way that's helpful and informative, while respecting important ethical boundaries.
+            type: reasoning.text
+          refusal: null
+          role: assistant
+        native_finish_reason: stop
+      created: 1760051228
+      id: gen-1760051228-zUtCCQbb0vkaM4UXZmcb
+      model: anthropic/claude-3.7-sonnet:thinking
+      object: chat.completion
+      provider: Google
+      usage:
+        completion_tokens: 402
+        prompt_tokens: 43
+        total_tokens: 445
+    status:
+      code: 200
+      message: OK
+version: 1

tests/models/cassettes/test_openrouter/test_openrouter_with_native_options.yaml

@@ -0,0 +1,82 @@
+interactions:
+- request:
+    headers:
+      accept:
+      - application/json
+      accept-encoding:
+      - gzip, deflate
+      connection:
+      - keep-alive
+      content-length:
+      - '193'
+      content-type:
+      - application/json
+      host:
+      - openrouter.ai
+    method: POST
+    parsed_body:
+      messages:
+      - content: Who are you
+        role: user
+      model: google/gemini-2.0-flash-exp:free
+      models:
+      - x-ai/grok-4
+      provider:
+        only:
+        - xai
+      stream: false
+      transforms:
+      - middle-out
+    uri: https://openrouter.ai/api/v1/chat/completions
+  response:
+    headers:
+      access-control-allow-origin:
+      - '*'
+      connection:
+      - keep-alive
+      content-length:
+      - '1067'
+      content-type:
+      - application/json
+      permissions-policy:
+      - payment=(self "https://checkout.stripe.com" "https://connect-js.stripe.com" "https://js.stripe.com" "https://*.js.stripe.com"
+        "https://hooks.stripe.com")
+      referrer-policy:
+      - no-referrer, strict-origin-when-cross-origin
+      transfer-encoding:
+      - chunked
+      vary:
+      - Accept-Encoding
+    parsed_body:
+      choices:
+      - finish_reason: stop
+        index: 0
+        logprobs: null
+        message:
+          content: |-
+            I'm Grok, a helpful and maximally truthful AI built by xAI. I'm not based on any other companies' models—instead, I'm inspired by the Hitchhiker's Guide to the Galaxy and JARVIS from Iron Man. My goal is to assist with questions, provide information, and maybe crack a joke or two along the way.
+
+            What can I help you with today?
+          reasoning: null
+          refusal: null
+          role: assistant
+        native_finish_reason: stop
+      created: 1759509677
+      id: gen-1759509677-MpJiZ3ZkiGU3lnbM8QKo
+      model: x-ai/grok-4
+      object: chat.completion
+      provider: xAI
+      system_fingerprint: fp_19e21a36c0
+      usage:
+        completion_tokens: 240
+        completion_tokens_details:
+          reasoning_tokens: 165
+        prompt_tokens: 687
+        prompt_tokens_details:
+          audio_tokens: 0
+          cached_tokens: 682
+        total_tokens: 927
+    status:
+      code: 200
+      message: OK
+version: 1