Skip to content

Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.

feat(bot): add MiniMax-M3 model support to minimax provider (#2368)

Octopus49e515667e0a
diff --git a/bot/README.md b/bot/README.md
index 30b23d6a8..ca91a0e5f 100644
--- a/bot/README.md
+++ b/bot/README.md
@@ -331,7 +331,7 @@ Provider configuration is read from OpenViking config (`vlm` section in `ov.conf
> - **Groq** provides free voice transcription via Whisper. If configured, Telegram voice messages will be automatically transcribed.
> - **Zhipu Coding Plan**: If you're on Zhipu's coding plan, set `"apiBase": "https://open.bigmodel.cn/api/coding/paas/v4"` in your zhipu provider config.
> - **MiniMax (Mainland China)**: If your API key is from MiniMax's mainland China platform (minimaxi.com), set `"apiBase": "https://api.minimaxi.com/v1"` in your minimax provider config.
-> - **MiniMax Recommended Models**: `MiniMax-M2.7` (peak performance) and `MiniMax-M2.7-highspeed` (faster, more agile). Configure with `"model": "MiniMax-M2.7"` in your agent config.
+> - **MiniMax Recommended Models**: `MiniMax-M3` (flagship, default), `MiniMax-M2.7` (peak performance) and `MiniMax-M2.7-highspeed` (faster, more agile). Configure with `"model": "MiniMax-M3"` in your agent config.
| Provider | Purpose | Get API Key |
|----------|---------|-------------|
diff --git a/bot/tests/test_minimax_provider.py b/bot/tests/test_minimax_provider.py
index 138f906c3..c5b177392 100644
--- a/bot/tests/test_minimax_provider.py
+++ b/bot/tests/test_minimax_provider.py
@@ -1,6 +1,6 @@
# Copyright (c) 2026 Beijing Volcano Engine Technology Co., Ltd.
# SPDX-License-Identifier: AGPL-3.0
-"""Tests for MiniMax provider support (MiniMax-M2.7, MiniMax-M2.7-highspeed)."""
+"""Tests for MiniMax provider support (MiniMax-M3, MiniMax-M2.7, MiniMax-M2.7-highspeed)."""
from urllib.parse import urlparse
@@ -28,6 +28,12 @@ class TestMiniMaxRegistry:
assert not spec.is_gateway
assert not spec.is_local
+ def test_minimax_m3_matched_by_keyword(self):
+ """MiniMax-M3 should be matched to the minimax ProviderSpec."""
+ spec = find_by_model("MiniMax-M3")
+ assert spec is not None, "MiniMax-M3 not matched to any provider"
+ assert spec.name == "minimax"
+
def test_minimax_m2_7_matched_by_keyword(self):
"""MiniMax-M2.7 should be matched to the minimax ProviderSpec."""
spec = find_by_model("MiniMax-M2.7")
@@ -42,7 +48,7 @@ class TestMiniMaxRegistry:
def test_minimax_keyword_is_case_insensitive(self):
"""Model name matching must be case-insensitive."""
- for model in ("minimax-m2.7", "MINIMAX-M2.7", "MiniMax-M2.7"):
+ for model in ("minimax-m2.7", "MINIMAX-M2.7", "MiniMax-M2.7", "MiniMax-m3"):
spec = find_by_model(model)
assert spec is not None, f"{model!r} not matched"
assert spec.name == "minimax"
@@ -71,6 +77,11 @@ class TestMiniMaxModelPrefixResolution:
model = f"{spec.litellm_prefix}/{model}"
return model
+ def test_m3_gets_minimax_prefix(self):
+ """MiniMax-M3 should be prefixed as minimax/MiniMax-M3."""
+ resolved = self._resolve_model("MiniMax-M3")
+ assert resolved == "minimax/MiniMax-M3"
+
def test_m2_7_gets_minimax_prefix(self):
"""MiniMax-M2.7 should be prefixed as minimax/MiniMax-M2.7."""
resolved = self._resolve_model("MiniMax-M2.7")
@@ -107,6 +118,18 @@ class TestMiniMaxSystemMessageHandling:
# LiteLLMProvider tests (model name after prefix resolution)
# ------------------------------------------------------------------ #
+ def test_litellm_system_message_merged_for_m3(self):
+ """System message is merged into the first user message for minimax/MiniMax-M3."""
+ messages = [
+ {"role": "system", "content": "You are a helpful assistant."},
+ {"role": "user", "content": "Hello!"},
+ ]
+ result = self._handle_system_litellm("minimax/MiniMax-M3", messages)
+ assert all(m["role"] != "system" for m in result), "System message not removed"
+ user_content = next(m["content"] for m in result if m["role"] == "user")
+ assert "You are a helpful assistant." in user_content
+ assert "Hello!" in user_content
+
def test_litellm_system_message_merged_for_m2_7(self):
"""System message is merged into the first user message for minimax/MiniMax-M2.7."""
messages = [
diff --git a/bot/vikingbot/providers/registry.py b/bot/vikingbot/providers/registry.py
index 758ff700f..22f3cb735 100644
--- a/bot/vikingbot/providers/registry.py
+++ b/bot/vikingbot/providers/registry.py
@@ -242,14 +242,15 @@ PROVIDERS: tuple[ProviderSpec, ...] = (
),
# MiniMax: needs "minimax/" prefix for LiteLLM routing.
# Uses OpenAI-compatible API at api.minimax.io/v1.
- # Recommended models: MiniMax-M2.7 (default), MiniMax-M2.7-highspeed (faster).
+ # Recommended models: MiniMax-M3 (default, flagship), MiniMax-M2.7 (alternative),
+ # MiniMax-M2.7-highspeed (faster).
# Note: MiniMax does not support system messages; they are merged into the first user message.
ProviderSpec(
name="minimax",
keywords=("minimax",),
env_key="MINIMAX_API_KEY",
display_name="MiniMax",
- litellm_prefix="minimax", # MiniMax-M2.7 → minimax/MiniMax-M2.7
+ litellm_prefix="minimax", # MiniMax-M3 → minimax/MiniMax-M3
skip_prefixes=("minimax/", "openrouter/"),
env_extras=(),
is_gateway=False,