raise HTTPStatusError so retry path skips, and keep tests green

Uses status_code gate (not is_error) and wraps the decoded x402 detail in
httpx.HTTPStatusError so _call_with_tee_retry correctly skips retries on
server-side failures. Adds request/url/headers to the test fake to match
the httpx Response contract.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

Author: balogh.adam@icloud.com

src/opengradient/client/llm.py

@@ -388,8 +388,12 @@ async def _request() -> TextGenerationOutput:
                 timeout=_REQUEST_TIMEOUT,
             )
             raw_body = await response.aread()
-            if response.is_error:
-                raise RuntimeError(_format_http_error(response, raw_body))
+            if response.status_code >= 400:
+                raise httpx.HTTPStatusError(
+                    _format_http_error(response, raw_body),
+                    request=response.request,
+                    response=response,
+                )
             result = json.loads(raw_body.decode())
 
             choices = result.get("choices")

tests/llm_test.py

@@ -7,7 +7,7 @@
 import json
 import ssl
 from contextlib import asynccontextmanager
-from typing import List
+from typing import Dict, List
 from unittest.mock import AsyncMock, MagicMock, patch
 
 import httpx
@@ -80,6 +80,9 @@ class _FakeResponse:
     def __init__(self, status_code: int, body: bytes):
         self.status_code = status_code
         self._body = body
+        self.headers: Dict[str, str] = {}
+        self.request = MagicMock()
+        self.url = "https://test.tee.server/v1/chat/completions"
 
     def raise_for_status(self):
         pass