feat: upgrade MiniMax default reasoning model to M3 (Tracer-Cloud#2698)

Bump MINIMAX_REASONING_MODEL from MiniMax-M2.7 to MiniMax-M3 to track
the latest MiniMax flagship. The high-speed classification / tool-call
slots stay on MiniMax-M2.7-highspeed since there is no M3 high-speed
variant yet.

- Update MINIMAX_REASONING_MODEL constant in app/config.py
- Update reasoning model references in docs/llm-providers.mdx
- Update affected unit tests in tests/test_config.py and
  tests/services/test_llm_client.py

Author: octo-patch

app/config.py

@@ -121,7 +121,7 @@ def get_environment() -> Environment:
 NVIDIA_TOOLCALL_MODEL = "meta/llama-3.1-8b-instruct"
 
 # MiniMax model constants
-MINIMAX_REASONING_MODEL = "MiniMax-M2.7"
+MINIMAX_REASONING_MODEL = "MiniMax-M3"
 MINIMAX_CLASSIFICATION_MODEL = "MiniMax-M2.7-highspeed"
 MINIMAX_TOOLCALL_MODEL = "MiniMax-M2.7-highspeed"
 

docs/llm-providers.mdx

@@ -15,7 +15,7 @@ OpenSRE is provider-agnostic: bring your own model. Selection is controlled by t
 | DeepSeek     | `deepseek`     | `DEEPSEEK_API_KEY`              | `deepseek-v4-pro`                             | `deepseek-v4-flash`                             |
 | Google Gemini| `gemini`       | `GEMINI_API_KEY`                | `gemini-3.1-pro-preview`                      | `gemini-3.1-flash-lite-preview`                 |
 | NVIDIA NIM   | `nvidia`       | `NVIDIA_API_KEY`                | `meta/llama-3.1-405b-instruct`                | `meta/llama-3.1-8b-instruct`                    |
-| MiniMax      | `minimax`      | `MINIMAX_API_KEY`               | `MiniMax-M2.7`                                | `MiniMax-M2.7-highspeed`                        |
+| MiniMax      | `minimax`      | `MINIMAX_API_KEY`               | `MiniMax-M3`                                  | `MiniMax-M2.7-highspeed`                        |
 | Amazon Bedrock| `bedrock`     | AWS IAM (`AWS_REGION`)          | `us.anthropic.claude-sonnet-4-6`              | `us.anthropic.claude-haiku-4-5-20251001-v1:0`   |
 | Ollama (local)| `ollama`     | None (local daemon)             | `llama3.2`                                    | `llama3.2`                                      |
 | OpenAI Codex CLI| `codex`    | `codex login` (CLI)             | Codex CLI default                             | Codex CLI default                               |
@@ -149,9 +149,9 @@ Uses NVIDIA's OpenAI-compatible API at `https://integrate.api.nvidia.com/v1`. Br
 export LLM_PROVIDER=minimax
 export MINIMAX_API_KEY=...
 # Optional override (single value applies to both slots if set):
-export MINIMAX_MODEL=MiniMax-M2.7
+export MINIMAX_MODEL=MiniMax-M3
 # Or per-slot:
-export MINIMAX_REASONING_MODEL=MiniMax-M2.7
+export MINIMAX_REASONING_MODEL=MiniMax-M3
 export MINIMAX_TOOLCALL_MODEL=MiniMax-M2.7-highspeed
 ```
 

tests/services/test_llm_client.py

@@ -380,7 +380,7 @@ def test_minimax_llm_client_reads_api_key_and_base_url(monkeypatch) -> None:
     monkeypatch.setattr(llm_client, "OpenAI", _FakeOpenAI)
 
     client = llm_client.OpenAILLMClient(
-        model="MiniMax-M2.7",
+        model="MiniMax-M3",
         base_url="https://api.minimax.io/v1",
         api_key_env="MINIMAX_API_KEY",
         temperature=1.0,
@@ -400,7 +400,7 @@ def test_minimax_llm_client_temperature_is_set(monkeypatch) -> None:
     monkeypatch.setattr(llm_client, "OpenAI", _FakeOpenAI)
 
     client = llm_client.OpenAILLMClient(
-        model="MiniMax-M2.7",
+        model="MiniMax-M3",
         base_url="https://api.minimax.io/v1",
         api_key_env="MINIMAX_API_KEY",
         temperature=1.0,

tests/test_config.py

@@ -82,7 +82,7 @@ def test_llm_settings_minimax_provider_accepted() -> None:
     )
     assert settings.provider == "minimax"
     assert settings.minimax_api_key == "mm-test-key"
-    assert settings.minimax_reasoning_model == "MiniMax-M2.7"
+    assert settings.minimax_reasoning_model == "MiniMax-M3"
     assert settings.minimax_toolcall_model == "MiniMax-M2.7-highspeed"