diff --git a/agent/auxiliary_client.py b/agent/auxiliary_client.py
index 360b487b6..d6af62210 100644
--- a/agent/auxiliary_client.py
+++ b/agent/auxiliary_client.py
@@ -265,9 +265,6 @@ _API_KEY_PROVIDER_AUX_MODELS_FALLBACK: Dict[str, str] = {
     "stepfun": "step-3.5-flash",
     "kimi-coding-cn": "kimi-k2-turbo-preview",
     "gmi": "google/gemini-3.1-flash-lite-preview",
-    "minimax": "MiniMax-M2.7",
-    "minimax-oauth": "MiniMax-M2.7-highspeed",
-    "minimax-cn": "MiniMax-M2.7",
     "anthropic": "claude-haiku-4-5-20251001",
     "opencode-zen": "gemini-3-flash",
     "opencode-go": "glm-5",
diff --git a/plugins/model-providers/minimax/__init__.py b/plugins/model-providers/minimax/__init__.py
index f29eb1aa0..6d77536ac 100644
--- a/plugins/model-providers/minimax/__init__.py
+++ b/plugins/model-providers/minimax/__init__.py
@@ -14,7 +14,7 @@ minimax = ProviderProfile(
     env_vars=("MINIMAX_API_KEY",),
     base_url="https://api.minimax.io/anthropic",
     auth_type="api_key",
-    default_aux_model="MiniMax-M2.7",
+    default_aux_model="MiniMax-M3",
 )
 
 minimax_cn = ProviderProfile(
@@ -24,7 +24,7 @@ minimax_cn = ProviderProfile(
     env_vars=("MINIMAX_CN_API_KEY",),
     base_url="https://api.minimaxi.com/anthropic",
     auth_type="api_key",
-    default_aux_model="MiniMax-M2.7",
+    default_aux_model="MiniMax-M3",
 )
 
 minimax_oauth = ProviderProfile(
@@ -37,7 +37,7 @@ minimax_oauth = ProviderProfile(
     env_vars=(),  # OAuth — tokens in auth.json, not env
     base_url="https://api.minimax.io/anthropic",
     auth_type="oauth_external",
-    default_aux_model="MiniMax-M2.7-highspeed",
+    default_aux_model="MiniMax-M2.7",
 )
 
 register_provider(minimax)
diff --git a/tests/agent/test_minimax_provider.py b/tests/agent/test_minimax_provider.py
index 73e8034dd..bed3c1166 100644
--- a/tests/agent/test_minimax_provider.py
+++ b/tests/agent/test_minimax_provider.py
@@ -141,12 +141,25 @@ class TestMinimaxThinkingSupport:
 
 
 class TestMinimaxAuxModel:
-    """Verify auxiliary model is standard (not highspeed) — now reads from profiles."""
+    """Verify auxiliary model is the current frontier standard (not highspeed).
+
+    As of M3's release (2026-06-01) the minimax / minimax-cn provider
+    profiles advertise ``MiniMax-M3`` as their ``default_aux_model`` (the
+    same model users see in ``_PROVIDER_MODELS["minimax"]`` and in the
+    user-facing ``model.default`` for a Token-Plan install).  The OAuth
+    / Coding Plan path sticks with M2.7 because M3 is not on that
+    tier — see ``test_minimax_profile.py`` for the per-provider split.
+
+    The historical concern this class guards is the #4082 / #6082
+    regression: the highspeed variant costs 2x with no model-quality
+    benefit, so we still assert that no aux choice contains the substring
+    ``"highspeed"``.
+    """
 
     def test_minimax_aux_is_standard(self):
         from agent.auxiliary_client import _get_aux_model_for_provider
-        assert _get_aux_model_for_provider("minimax") == "MiniMax-M2.7"
-        assert _get_aux_model_for_provider("minimax-cn") == "MiniMax-M2.7"
+        assert _get_aux_model_for_provider("minimax") == "MiniMax-M3"
+        assert _get_aux_model_for_provider("minimax-cn") == "MiniMax-M3"
 
     def test_minimax_aux_not_highspeed(self):
         from agent.auxiliary_client import _get_aux_model_for_provider
diff --git a/tests/plugins/model_providers/test_minimax_profile.py b/tests/plugins/model_providers/test_minimax_profile.py
new file mode 100644
index 000000000..495c125ce
--- /dev/null
+++ b/tests/plugins/model_providers/test_minimax_profile.py
@@ -0,0 +1,119 @@
+"""Unit tests for the MiniMax provider profile.
+
+Three MiniMax provider profiles (`minimax` direct API, `minimax-cn` China direct
+API, `minimax-oauth` browser OAuth) all advertise a `default_aux_model` on
+their `ProviderProfile`. The previous M2.7 / M2.7-highspeed values were
+stale relative to the current frontier model (M3, released 2026-06-01) and
+inconsistent with the `_PROVIDER_MODELS["minimax"]` catalog top entry in
+`hermes_cli/models.py`.
+
+This file pins the new defaults so the choice is reviewable and any future
+revert shows up in a failing test rather than silent behavior drift.
+
+Refs:
+  - Issue #36196: M3 support request
+  - PR #36205 (closed unmerged): Csrayz's M3 + 1M context work
+  - PR #36212 (open): adds M3 to `_PROVIDER_MODELS["minimax"]` catalog
+  - PR #6082: M2.7-highspeed → M2.7 for aux model (half-price fix)
+  - Commit 773a0faca: same profile-layer fix pattern for `deepseek`
+"""
+
+from __future__ import annotations
+
+import pytest
+
+
+@pytest.fixture(params=["minimax", "minimax-cn", "minimax-oauth"])
+def minimax_profile(request):
+    """Resolve each registered MiniMax profile.
+
+    Going through ``providers.get_provider_profile`` keeps the test honest —
+    if someone later replaces the registered class with a plain
+    ``ProviderProfile``, every assertion below collapses.
+    """
+    import model_tools  # noqa: F401  -- triggers plugin discovery
+    import providers
+
+    profile = providers.get_provider_profile(request.param)
+    assert profile is not None, f"{request.param} provider profile must be registered"
+    return profile, request.param
+
+
+class TestMinimaxAuxModelM3:
+    """MiniMax profile aux model is the new frontier M3, not the stale M2.7.
+
+    The catalog top entry is ``MiniMax-M3`` (hermes_cli/models.py:298) and
+    the user-facing ``model.default`` for a Token-Plan install is M3, so
+    pinning the aux default to the same model keeps the runtime consistent
+    (same auth, same billing pool, same rate limits, no surprise 2x-cost
+    highspeed variant). M3 was released 2026-06-01 — picking it as the
+    aux default matches the forward-looking catalog order rather than the
+    pre-M3 era.
+    """
+
+    @pytest.mark.parametrize(
+        "provider_id,expected",
+        [
+            ("minimax", "MiniMax-M3"),
+            ("minimax-cn", "MiniMax-M3"),
+            # minimax-oauth sticks with M2.7: the OAuth / Coding Plan
+            # tier historically used -highspeed (PR #6082 collapsed that
+            # to plain M2.7 to avoid the 2x TPS surcharge). M3 is not on
+            # the OAuth/Coding Plan tier per platform docs as of this PR,
+            # so the safe choice is the cheapest generally-available
+            # M2.7 — matching PR #6082's intent.
+            ("minimax-oauth", "MiniMax-M2.7"),
+        ],
+    )
+    def test_profile_advertises_expected_aux_model(
+        self, provider_id, expected
+    ):
+        import model_tools  # noqa: F401
+        import providers
+
+        profile = providers.get_provider_profile(provider_id)
+        assert profile is not None
+        assert profile.default_aux_model == expected, (
+            f"{provider_id} default_aux_model drifted to "
+            f"{profile.default_aux_model!r}, expected {expected!r}"
+        )
+
+    def test_consumer_api_returns_non_empty_for_each_provider(self, minimax_profile):
+        from agent.auxiliary_client import _get_aux_model_for_provider
+
+        profile, provider_id = minimax_profile
+        resolved = _get_aux_model_for_provider(provider_id)
+        assert resolved != "", (
+            f"_get_aux_model_for_provider({provider_id!r}) returned empty — "
+            "the 'No auxiliary LLM provider configured' warning will fire on "
+            f"every {provider_id} session even though the profile advertises "
+            f"default_aux_model={profile.default_aux_model!r}"
+        )
+        assert resolved == profile.default_aux_model, (
+            f"_get_aux_model_for_provider({provider_id!r}) returned "
+            f"{resolved!r} but profile advertises {profile.default_aux_model!r} "
+            "— the consumer API and the profile have drifted out of sync"
+        )
+
+
+class TestMinimaxAuxModelNotHighspeed:
+    """Regression guard against re-introducing the M2.7-highspeed aux default.
+
+    PR #6082 collapsed the highspeed aux choice to plain M2.7 because the
+    highspeed variant costs 2x with no real benefit for compression / vision /
+    session-search aux tasks. None of the three MiniMax profiles should
+    silently re-introduce that 2x-cost path.
+    """
+
+    @pytest.mark.parametrize("provider_id", ["minimax", "minimax-cn", "minimax-oauth"])
+    def test_default_aux_model_is_not_highspeed(self, provider_id):
+        import model_tools  # noqa: F401
+        import providers
+
+        profile = providers.get_provider_profile(provider_id)
+        assert profile is not None
+        assert "highspeed" not in profile.default_aux_model.lower(), (
+            f"{provider_id} default_aux_model={profile.default_aux_model!r} "
+            "is a -highspeed variant — that costs 2x for the same model and "
+            "broke #4082 the first time. Revert to plain M2.7 or M3."
+        )