feat: use endpoint metadata for custom model context and pricing (#1906)

* perf: cache base_url.lower() via property, consolidate triple load_config(), hoist set constant

run_agent.py:
- Add base_url property that auto-caches _base_url_lower on every
  assignment, eliminating 12+ redundant .lower() calls per API cycle
  across __init__, _build_api_kwargs, _supports_reasoning_extra_body,
  and the main conversation loop
- Consolidate three separate load_config() disk reads in __init__
  (memory, skills, compression) into a single call, reusing the
  result dict for all three config sections

model_tools.py:
- Hoist _READ_SEARCH_TOOLS set to module level (was rebuilt inside
  handle_function_call on every tool invocation)

* Use endpoint metadata for custom model context and pricing

---------

Co-authored-by: kshitij <82637225+kshitijk4poor@users.noreply.github.com>

tests/agent/test_model_metadata.py

@@ -188,6 +188,36 @@ class TestGetModelContextLength:
             result = get_model_context_length("custom/model")
             assert result == CONTEXT_PROBE_TIERS[0]
 
+    @patch("agent.model_metadata.fetch_model_metadata")
+    @patch("agent.model_metadata.fetch_endpoint_model_metadata")
+    def test_custom_endpoint_metadata_beats_fuzzy_default(self, mock_endpoint_fetch, mock_fetch):
+        mock_fetch.return_value = {}
+        mock_endpoint_fetch.return_value = {
+            "zai-org/GLM-5-TEE": {"context_length": 65536}
+        }
+
+        result = get_model_context_length(
+            "zai-org/GLM-5-TEE",
+            base_url="https://llm.chutes.ai/v1",
+            api_key="test-key",
+        )
+
+        assert result == 65536
+
+    @patch("agent.model_metadata.fetch_model_metadata")
+    @patch("agent.model_metadata.fetch_endpoint_model_metadata")
+    def test_custom_endpoint_without_metadata_skips_name_based_default(self, mock_endpoint_fetch, mock_fetch):
+        mock_fetch.return_value = {}
+        mock_endpoint_fetch.return_value = {}
+
+        result = get_model_context_length(
+            "zai-org/GLM-5-TEE",
+            base_url="https://llm.chutes.ai/v1",
+            api_key="test-key",
+        )
+
+        assert result == CONTEXT_PROBE_TIERS[0]
+
 
 # =========================================================================
 # fetch_model_metadata — caching, TTL, slugs, failures
@@ -258,6 +288,25 @@ class TestFetchModelMetadata:
         assert "anthropic/claude-3.5-sonnet" in result
         assert result["anthropic/claude-3.5-sonnet"]["context_length"] == 200000
 
+    @patch("agent.model_metadata.requests.get")
+    def test_provider_prefixed_models_get_bare_aliases(self, mock_get):
+        self._reset_cache()
+        mock_response = MagicMock()
+        mock_response.json.return_value = {
+            "data": [{
+                "id": "provider/test-model",
+                "context_length": 123456,
+                "name": "Provider: Test Model",
+            }]
+        }
+        mock_response.raise_for_status = MagicMock()
+        mock_get.return_value = mock_response
+
+        result = fetch_model_metadata(force_refresh=True)
+
+        assert result["provider/test-model"]["context_length"] == 123456
+        assert result["test-model"]["context_length"] == 123456
+
     @patch("agent.model_metadata.requests.get")
     def test_ttl_expiry_triggers_refetch(self, mock_get):
         """Cache expires after _MODEL_CACHE_TTL seconds."""

tests/agent/test_usage_pricing.py

@@ -99,3 +99,27 @@ def test_estimate_usage_cost_refuses_cache_pricing_without_official_cache_rate(m
     )
 
     assert result.status == "unknown"
+
+
+def test_custom_endpoint_models_api_pricing_is_supported(monkeypatch):
+    monkeypatch.setattr(
+        "agent.usage_pricing.fetch_endpoint_model_metadata",
+        lambda base_url, api_key=None: {
+            "zai-org/GLM-5-TEE": {
+                "pricing": {
+                    "prompt": "0.0000005",
+                    "completion": "0.000002",
+                }
+            }
+        },
+    )
+
+    entry = get_pricing_entry(
+        "zai-org/GLM-5-TEE",
+        provider="custom",
+        base_url="https://llm.chutes.ai/v1",
+        api_key="test-key",
+    )
+
+    assert float(entry.input_cost_per_million) == 0.5
+    assert float(entry.output_cost_per_million) == 2.0