test(providers): cover reasoning_effort="none" and gemma auto-routing

- Anthropic: "none" must not enable extended thinking
- Azure: "none" must not suppress temperature or inject reasoning body
- DeepSeek/DashScope/Kimi: "none" sends thinking disabled, skips reasoning_effort field
- Gemini: gemma keyword enables auto-routing for gemma models

tests/providers/test_anthropic_thinking.py

@@ -83,3 +83,10 @@ def test_opus_4_7_omits_temperature_none() -> None:
     kw = _build(_make_provider("claude-opus-4-7"), None)
     assert "temperature" not in kw
     assert "thinking" not in kw
+
+
+def test_reasoning_effort_string_none_does_not_enable_thinking() -> None:
+    """reasoning_effort='none' must not enable thinking — treated same as disabled."""
+    kw = _build(_make_provider(), "none")
+    assert "thinking" not in kw
+    assert kw["temperature"] == 0.7

tests/providers/test_azure_openai_provider.py

@@ -78,6 +78,11 @@ def test_supports_temperature_with_reasoning_effort():
     assert AzureOpenAIProvider._supports_temperature("gpt-4o", reasoning_effort="medium") is False
 
 
+def test_supports_temperature_with_reasoning_effort_none_string():
+    """reasoning_effort='none' must NOT suppress temperature — it means thinking is off."""
+    assert AzureOpenAIProvider._supports_temperature("gpt-4o", reasoning_effort="none") is True
+
+
 # ---------------------------------------------------------------------------
 # _build_body — Responses API body construction
 # ---------------------------------------------------------------------------
@@ -131,6 +136,16 @@ def test_build_body_with_reasoning():
     assert "temperature" not in body
 
 
+def test_build_body_reasoning_effort_none_string_omits_reasoning():
+    """reasoning_effort='none' must not inject a reasoning body and must allow temperature."""
+    provider = AzureOpenAIProvider(api_key="k", api_base="https://r.com", default_model="gpt-4o")
+    body = provider._build_body(
+        [{"role": "user", "content": "hi"}], None, "gpt-4o", 4096, 0.7, "none", None,
+    )
+    assert "reasoning" not in body
+    assert body["temperature"] == 0.7
+
+
 def test_build_body_image_conversion():
     """image_url content blocks should be converted to input_image."""
     provider = AzureOpenAIProvider(api_key="k", api_base="https://r.com", default_model="gpt-4o")

tests/providers/test_litellm_kwargs.py

@@ -121,6 +121,14 @@ def test_openrouter_spec_is_gateway() -> None:
     assert spec.default_api_base == "https://openrouter.ai/api/v1"
 
 
+def test_gemma_routes_to_gemini_provider() -> None:
+    """gemma models (e.g. gemma-3-27b-it) must auto-route to Gemini when GEMINI_API_KEY is set.
+    Users running gemma via the Gemini API endpoint expect automatic provider detection."""
+    spec = find_by_name("gemini")
+    assert spec is not None
+    assert "gemma" in spec.keywords
+
+
 def test_openrouter_sets_default_attribution_headers() -> None:
     spec = find_by_name("openrouter")
     with patch("nanobot.providers.openai_compat_provider.AsyncOpenAI") as MockClient:
@@ -1050,3 +1058,46 @@ def test_kimi_k2_thinking_series_no_thinking_injection() -> None:
     """kimi-k2-thinking series models must NOT receive extra_body.thinking."""
     kw = _build_kwargs_for("moonshot", "kimi-k2-thinking", reasoning_effort="high")
     assert "extra_body" not in kw
+
+
+# ---------------------------------------------------------------------------
+# reasoning_effort="none" — treated as thinking disabled 
+# ---------------------------------------------------------------------------
+
+def test_deepseek_thinking_disabled_for_none_string() -> None:
+    """reasoning_effort='none' must send thinking.type=disabled and skip reasoning_effort field."""
+    kw = _build_kwargs_for("deepseek", "deepseek-v4-pro", reasoning_effort="none")
+    assert kw.get("extra_body") == {"thinking": {"type": "disabled"}}
+    assert "reasoning_effort" not in kw
+
+
+def test_kimi_k25_thinking_disabled_for_none_string() -> None:
+    """reasoning_effort='none' maps to thinking disabled for kimi-k2.5."""
+    kw = _build_kwargs_for("moonshot", "kimi-k2.5", reasoning_effort="none")
+    assert kw.get("extra_body") == {"thinking": {"type": "disabled"}}
+
+
+def test_dashscope_thinking_disabled_for_none_string() -> None:
+    """reasoning_effort='none' disables thinking and must not emit reasoning_effort on DashScope."""
+    kw = _build_kwargs_for("dashscope", "qwen3.6-plus", reasoning_effort="none")
+    assert kw.get("extra_body") == {"enable_thinking": False}
+    assert "reasoning_effort" not in kw
+
+
+def test_deepseek_no_backfill_when_reasoning_effort_none_string() -> None:
+    """reasoning_effort='none' must NOT trigger reasoning_content backfill (thinking inactive)."""
+    spec = find_by_name("deepseek")
+    with patch("nanobot.providers.openai_compat_provider.AsyncOpenAI"):
+        p = OpenAICompatProvider(api_key="k", default_model="deepseek-v4-pro", spec=spec)
+    messages = [
+        {"role": "user", "content": "hi"},
+        {"role": "assistant", "content": "ok"},
+        {"role": "user", "content": "continue"},
+    ]
+    kw = p._build_kwargs(
+        messages=list(messages), tools=None, model="deepseek-v4-pro",
+        max_tokens=1024, temperature=0.7,
+        reasoning_effort="none", tool_choice=None,
+    )
+    assistant = kw["messages"][1]
+    assert "reasoning_content" not in assistant