feat(kimi): add k2.6-code-preview support across kimi-coding, kimi-coding-cn, and moonshot

agent/model_metadata.py

@@ -154,6 +154,7 @@ def _strip_provider_prefix(model: str) -> str:
     "grok-2": 131072,           # grok-2, grok-2-1212, grok-2-latest
     "grok": 131072,             # catch-all (grok-beta, unknown grok-*)
     # Kimi
+    "k2.6-code-preview": 256000,
     "kimi": 262144,
     # Arcee
     "trinity": 262144,

hermes_cli/model_switch.py

@@ -500,6 +500,25 @@ def switch_model(
             )
 
         target_provider = pdef.id
+        # Some providers have a models.dev ID that differs from the Hermes
+        # internal ID (e.g. models.dev uses "kimi-for-coding" but Hermes uses
+        # "kimi-coding" everywhere else). Prefer the Hermes internal ID when
+        # the user explicitly named a provider we know natively.
+        try:
+            from hermes_cli.auth import PROVIDER_REGISTRY
+            _HERMES_ALIASES = {
+                "kimi": "kimi-coding",
+                "kimi-for-coding": "kimi-coding",
+                "moonshot": "kimi-coding",
+                "kimi-cn": "kimi-coding-cn",
+                "moonshot-cn": "kimi-coding-cn",
+            }
+            _raw = explicit_provider.strip().lower()
+            _canonical_explicit = _HERMES_ALIASES.get(_raw, _raw)
+            if _canonical_explicit in PROVIDER_REGISTRY:
+                target_provider = _canonical_explicit
+        except Exception:
+            pass
 
         # If no model specified, try auto-detect from endpoint
         if not new_model:

hermes_cli/models.py

@@ -157,6 +157,7 @@ def _codex_curated_models() -> list[str]:
     ],
     "kimi-coding": [
         "kimi-for-coding",
+        "k2.6-code-preview",
         "kimi-k2.5",
         "kimi-k2-thinking",
         "kimi-k2-thinking-turbo",
@@ -165,11 +166,13 @@ def _codex_curated_models() -> list[str]:
     ],
     "kimi-coding-cn": [
         "kimi-k2.5",
+        "k2.6-code-preview",
         "kimi-k2-thinking",
         "kimi-k2-turbo-preview",
         "kimi-k2-0905-preview",
     ],
     "moonshot": [
+        "k2.6-code-preview",
         "kimi-k2.5",
         "kimi-k2-thinking",
         "kimi-k2-turbo-preview",
@@ -1949,6 +1952,17 @@ def validate_requested_model(
             # listing (e.g. Z.AI Pro/Max plans can use glm-5 on coding
             # endpoints even though it's not in /models).  Warn but allow.
 
+            # Trust our static curated list when the live API is incomplete
+            # (e.g. Kimi Coding Plan only exposes kimi-for-coding).
+            static_models = _PROVIDER_MODELS.get(normalized, [])
+            if requested_for_lookup in static_models:
+                return {
+                    "accepted": True,
+                    "persist": True,
+                    "recognized": True,
+                    "message": None,
+                }
+
             # Auto-correct if the top match is very similar (e.g. typo)
             auto = get_close_matches(requested_for_lookup, api_models, n=1, cutoff=0.9)
             if auto:
@@ -2015,6 +2029,18 @@ def validate_requested_model(
             pass  # Fall through to generic warning
 
     provider_label = _PROVIDER_LABELS.get(normalized, normalized)
+
+    # Trust our static curated list even when the live API is unreachable.
+    # If we hand-picked the model, there's no value in warning the user.
+    static_models = _PROVIDER_MODELS.get(normalized, [])
+    if requested_for_lookup in static_models:
+        return {
+            "accepted": True,
+            "persist": True,
+            "recognized": True,
+            "message": None,
+        }
+
     return {
         "accepted": True,
         "persist": True,

hermes_cli/setup.py

@@ -97,8 +97,8 @@ def _supports_same_provider_pool_setup(provider: str) -> bool:
         "gemma-4-31b-it", "gemma-4-26b-it",
     ],
     "zai": ["glm-5.1", "glm-5", "glm-4.7", "glm-4.5", "glm-4.5-flash"],
-    "kimi-coding": ["kimi-k2.5", "kimi-k2-thinking", "kimi-k2-turbo-preview"],
-    "kimi-coding-cn": ["kimi-k2.5", "kimi-k2-thinking", "kimi-k2-turbo-preview"],
+    "kimi-coding": ["k2.6-code-preview", "kimi-k2.5", "kimi-k2-thinking", "kimi-k2-turbo-preview"],
+    "kimi-coding-cn": ["k2.6-code-preview", "kimi-k2.5", "kimi-k2-thinking", "kimi-k2-turbo-preview"],
     "arcee": ["trinity-large-thinking", "trinity-large-preview", "trinity-mini"],
     "minimax": ["MiniMax-M2.7", "MiniMax-M2.5", "MiniMax-M2.1", "MiniMax-M2"],
     "minimax-cn": ["MiniMax-M2.7", "MiniMax-M2.5", "MiniMax-M2.1", "MiniMax-M2"],

tests/hermes_cli/test_model_validation.py

@@ -163,7 +163,7 @@ def test_case_insensitive(self):
 class TestProviderLabel:
     def test_known_labels_and_auto(self):
         assert provider_label("anthropic") == "Anthropic"
-        assert provider_label("kimi") == "Kimi / Moonshot"
+        assert provider_label("kimi") == "Kimi / Kimi Coding Plan"
         assert provider_label("copilot") == "GitHub Copilot"
         assert provider_label("copilot-acp") == "GitHub Copilot ACP"
         assert provider_label("auto") == "Auto"
@@ -453,6 +453,16 @@ def test_dissimilar_model_shows_suggestions_not_autocorrect(self):
         assert result.get("corrected_model") is None
         assert "not found" in result["message"]
 
+    def test_static_list_trusted_when_api_is_incomplete(self):
+        """If the live API omits a model we curate statically, trust the static list."""
+        # Simulate Kimi Coding Plan's /models endpoint, which only exposes
+        # kimi-for-coding even though newer models like k2.6-code-preview work.
+        with patch("hermes_cli.models.fetch_api_models", return_value=["kimi-for-coding"]):
+            result = validate_requested_model("k2.6-code-preview", "kimi-coding")
+        assert result["accepted"] is True
+        assert result["recognized"] is True
+        assert result["message"] is None
+
 
 # -- validate — API unreachable — accept and persist everything ----------------