feat: add MiniMax M2.7 to validated catalog

AGENTS.md

@@ -36,9 +36,9 @@ Current honest state:
   validated curated default, with installer-managed
   `limit.output = 8192` for Kilo `7.2.0` compatibility
 - the managed GonkaGate provider config now writes the validated
-  chat-completions catalog, currently Kimi K2.6 plus Qwen3 235B A22B Instruct
-  2507 FP8, so Kilo's OpenCode-style `/models` picker can switch between
-  validated GonkaGate models
+  chat-completions catalog, currently Kimi K2.6, Qwen3 235B A22B Instruct
+  2507 FP8, and MiniMax M2.7, so Kilo's OpenCode-style `/models` picker can
+  switch between validated GonkaGate models
 - Kilo detection now accepts `@kilocode/cli >=7.2.0` without pre-blocking
   future Kilo releases, while the audited compatibility baseline remains
   `@kilocode/cli@7.2.0`
@@ -136,8 +136,8 @@ These are implementation facts today:
   orchestration
 - `src/constants/models.ts` now exposes Kimi K2.6 as the recommended validated
   production default plus the validated Qwen3 235B A22B Instruct 2507 FP8
-  catalog entry, each with installer-managed `limit.output = 8192` in the
-  written Kilo provider config
+  and MiniMax M2.7 catalog entries, each with installer-managed
+  `limit.output = 8192` in the written Kilo provider config
 - `docs/specs/kilo-setup-prd/spec.md` is the copied Kilo setup PRD
 - `docs/release-readiness.md` records the current production-readiness audit
 - `.github/workflows/release-please.yml` and `.github/workflows/publish.yml`

CHANGELOG.md

@@ -4,7 +4,7 @@
 
 - Wrote the validated GonkaGate chat-completions catalog into the managed
   provider config so Kilo's OpenCode-style `/models` picker can switch between
-  Kimi K2.6 and Qwen3 235B A22B Instruct 2507 FP8.
+  Kimi K2.6, Qwen3 235B A22B Instruct 2507 FP8, and MiniMax M2.7.
 - Accepted Kilo versions at or above `@kilocode/cli >=7.2.0` instead of
   pre-blocking future Kilo releases with the previous `7.2.0`-only gate.
 - Kept local verification blockers and mismatches visible in rollback output

README.md

@@ -75,8 +75,9 @@ Current public baseline:
 - curated default:
   `moonshotai/Kimi-K2.6`
 - managed validated catalog for Kilo's OpenCode-style `/models` picker:
-  `moonshotai/Kimi-K2.6` and
-  `qwen/qwen3-235b-a22b-instruct-2507-fp8`
+  `moonshotai/Kimi-K2.6`,
+  `qwen/qwen3-235b-a22b-instruct-2507-fp8`, and
+  `minimaxai/minimax-m2.7`
 - installer-managed `limit.output = 8192` for Kilo compatibility
 - no native Windows production claim yet
 
@@ -207,8 +208,9 @@ This repository intentionally stays narrow today:
 - current curated default:
   `moonshotai/Kimi-K2.6`
 - validated GonkaGate models exposed to Kilo's `/models` picker:
-  `moonshotai/Kimi-K2.6` and
-  `qwen/qwen3-235b-a22b-instruct-2507-fp8`
+  `moonshotai/Kimi-K2.6`,
+  `qwen/qwen3-235b-a22b-instruct-2507-fp8`, and
+  `minimaxai/minimax-m2.7`
 - real-path Kilo verification is not the production default
 - native Windows production support is not claimed yet
 - future Kilo releases are not pre-blocked by version, but observed

docs/README.md

@@ -24,6 +24,7 @@ Current contract documents:
 
 This repository ships the Kilo installer runtime with a validated curated
 GonkaGate chat-completions catalog, Kimi K2.6 as the recommended default,
-installer-managed `limit.output = 8192` for Kilo compatibility, and a minimum
-accepted Kilo floor of `@kilocode/cli >=7.2.0` without a preset upper version
-bound.
+MiniMax M2.7 and Qwen3 235B A22B Instruct 2507 FP8 as additional validated
+catalog entries, installer-managed `limit.output = 8192` for Kilo
+compatibility, and a minimum accepted Kilo floor of `@kilocode/cli >=7.2.0`
+without a preset upper version bound.

docs/gonkagate-x-kilo.md

@@ -131,9 +131,11 @@ Right now the public default is deliberately small:
 - minimum Kilo floor: `@kilocode/cli >=7.2.0`
 - recommended validated model: `moonshotai/Kimi-K2.6`
 - validated catalog exposed to Kilo's OpenCode-style `/models` picker:
-  `moonshotai/Kimi-K2.6` and
-  `qwen/qwen3-235b-a22b-instruct-2507-fp8`
-- managed limits: `limit.context = 262144`, `limit.output = 8192`
+  `moonshotai/Kimi-K2.6`,
+  `qwen/qwen3-235b-a22b-instruct-2507-fp8`, and
+  `minimaxai/minimax-m2.7`
+- managed limits: `limit.output = 8192` for all validated entries;
+  `limit.context = 262144` for Kimi/Qwen and `204800` for MiniMax
 
 We are treating model support as a curated list, not as a vague "it probably
 works" promise.

docs/how-it-works.md

@@ -33,9 +33,10 @@ Current public limit:
   `moonshotai/Kimi-K2.6` with `limit.context = 262144` and
   `limit.output = 8192`
 - the written provider config includes the validated chat-completions model
-  catalog, currently `moonshotai/Kimi-K2.6` and
-  `qwen/qwen3-235b-a22b-instruct-2507-fp8`, so Kilo's OpenCode-style
-  `/models` picker can switch between validated GonkaGate models
+  catalog, currently `moonshotai/Kimi-K2.6`,
+  `qwen/qwen3-235b-a22b-instruct-2507-fp8`, and
+  `minimaxai/minimax-m2.7`, so Kilo's OpenCode-style `/models` picker can
+  switch between validated GonkaGate models
 - broader model claims still require additional proof before models should be
   added to the validated catalog
 

docs/release-readiness.md

@@ -20,10 +20,12 @@ docs, and tests for these facts:
 - curated public default:
   `moonshotai/Kimi-K2.6`
 - validated provider catalog exposed to Kilo's OpenCode-style `/models` picker:
-  `moonshotai/Kimi-K2.6` and
-  `qwen/qwen3-235b-a22b-instruct-2507-fp8`
+  `moonshotai/Kimi-K2.6`,
+  `qwen/qwen3-235b-a22b-instruct-2507-fp8`, and
+  `minimaxai/minimax-m2.7`
 - curated model limits for the validated catalog:
-  `limit.context = 262144`; `limit.output = 8192`
+  `limit.output = 8192` for all validated entries; `limit.context = 262144`
+  for Kimi/Qwen and `204800` for MiniMax
 - managed secret path: `~/.gonkagate/kilo/api-key`
 - project scope stays secret-free and still depends on a compatible
   user-level `provider.gonkagate` definition on each machine
@@ -46,6 +48,10 @@ Moonshot metadata checked on 2026-04-29:
 - the April 14, 2026 Kilo compatibility spike captured
   `qwen/qwen3-235b-a22b-instruct-2507-fp8` as a validated
   chat-completions model with a 262K context window and `maxTokens` 8192
+- NVIDIA NIM metadata checked on 2026-05-29 lists
+  `minimaxai/minimax-m2.7` as MiniMax M2.7 with a 204,800 input context
+  length; the package keeps the same installer-managed `limit.output = 8192`
+  compatibility clamp for this catalog entry
 - npm registry metadata checked on 2026-04-29 showed `@kilocode/cli` patch
   releases in the `7.2.x` line, including `7.2.14`, with both `kilo` and
   `kilocode` binaries still exposed by the wrapper package
@@ -76,4 +82,5 @@ repository-only pass:
 Those items do not change the current package contract: the publishable surface
 remains the minimum Kilo `7.2.0` floor without an upper version bound, the
 validated GonkaGate chat-completions catalog with Kimi as the recommended
-default, and the current non-Windows verification policy documented above.
+default, Qwen3 and MiniMax as additional validated entries, and the current
+non-Windows verification policy documented above.

docs/troubleshooting.md

@@ -17,14 +17,15 @@ If you see a config error mentioning
 you are likely running a stale build that predates the required Kilo 7.2.0
 output-limit fix.
 
-## Why Does `/models` Show Only One GonkaGate Model?
+## Why Does `/models` Show Too Few GonkaGate Models?
 
 Current builds write the validated GonkaGate chat-completions catalog into
 `provider.gonkagate.models`, so Kilo's OpenCode-style `/models` picker should
-show both `moonshotai/Kimi-K2.6` and
-`qwen/qwen3-235b-a22b-instruct-2507-fp8`.
+show `moonshotai/Kimi-K2.6`,
+`qwen/qwen3-235b-a22b-instruct-2507-fp8`, and
+`minimaxai/minimax-m2.7`.
 
-If `/models` still shows only one GonkaGate model, rerun the installer from a
+If `/models` still shows fewer GonkaGate models, rerun the installer from a
 current build so the user-level `provider.gonkagate` definition is refreshed.
 Also check for a higher-precedence Kilo layer such as project config,
 `KILO_CONFIG`, `KILO_CONFIG_DIR`, or `KILO_CONFIG_CONTENT` defining its own

docs/user-guide.md

@@ -26,9 +26,11 @@ Today the public package is intentionally narrow:
 - current curated default:
   `moonshotai/Kimi-K2.6`
 - validated GonkaGate catalog for Kilo's OpenCode-style `/models` picker:
-  `moonshotai/Kimi-K2.6` and
-  `qwen/qwen3-235b-a22b-instruct-2507-fp8`
-- managed model limits: `limit.context = 262144`, `limit.output = 8192`
+  `moonshotai/Kimi-K2.6`,
+  `qwen/qwen3-235b-a22b-instruct-2507-fp8`, and
+  `minimaxai/minimax-m2.7`
+- managed model limits: `limit.output = 8192` for all validated entries;
+  `limit.context = 262144` for Kimi/Qwen and `204800` for MiniMax
 - no native Windows production claim yet
 
 ## Before You Run It

src/constants/models.ts

@@ -117,6 +117,24 @@ export const CURATED_MODEL_REGISTRY = Object.freeze({
     transport: "chat_completions",
     validationStatus: "validated",
   },
+  "minimax-m2.7": {
+    adapterPackage: "@ai-sdk/openai-compatible",
+    displayName: "MiniMax M2.7",
+    limits: {
+      context: 204800,
+      output: 8192,
+    },
+    modelId: "minimaxai/minimax-m2.7",
+    recommended: false,
+    runtimeCompatibility: {
+      notes: [
+        "NVIDIA NIM metadata lists minimaxai/minimax-m2.7 as MiniMax M2.7 with a 204,800 input context length.",
+        "The installer writes limit.output = 8192 as the existing conservative Kilo 7.2.0 compatibility clamp for validated catalog entries.",
+      ],
+    },
+    transport: "chat_completions",
+    validationStatus: "validated",
+  },
 } as const satisfies CuratedModelRegistry);
 
 function toCuratedModelRecord<

src/install/README.md

@@ -20,6 +20,6 @@ Start here:
   durable/current-session Kilo proof path with redacted diagnostics
 
 The current stock runtime ships a validated curated Kimi K2.6 default plus the
-validated Qwen3 235B A22B Instruct 2507 FP8 catalog entry. Keep Kilo-specific
-behavior here and do not import OpenCode env vars or config targets as runtime
-defaults.
+validated Qwen3 235B A22B Instruct 2507 FP8 and MiniMax M2.7 catalog entries.
+Keep Kilo-specific behavior here and do not import OpenCode env vars or config
+targets as runtime defaults.

test/install/models.test.ts

@@ -17,6 +17,7 @@ test("the default curated model is the shipped validated Kimi public default", (
   assert.deepEqual(getValidatedModelKeys(), [
     "kimi-k2.6",
     "qwen3-235b-a22b-instruct-2507-fp8",
+    "minimax-m2.7",
   ]);
   assert.equal(getRecommendedProductionDefaultModel()?.key, "kimi-k2.6");
 });

test/package-contract.test.ts

@@ -134,7 +134,8 @@ test("constants use Kilo-specific defaults and leave OpenCode env vars out", ()
 
 test("curated model registry exposes Kimi as the shipped validated default", () => {
   const kimi = CURATED_MODEL_REGISTRY["kimi-k2.6"];
-  const model = CURATED_MODEL_REGISTRY["qwen3-235b-a22b-instruct-2507-fp8"];
+  const qwen = CURATED_MODEL_REGISTRY["qwen3-235b-a22b-instruct-2507-fp8"];
+  const minimax = CURATED_MODEL_REGISTRY["minimax-m2.7"];
 
   assert.equal(kimi.adapterPackage, "@ai-sdk/openai-compatible");
   assert.deepEqual(kimi.limits, { context: 262144, output: 8192 });
@@ -143,27 +144,38 @@ test("curated model registry exposes Kimi as the shipped validated default", ()
   assert.equal(kimi.transport, "chat_completions");
   assert.equal(kimi.validationStatus, "validated");
 
-  assert.equal(model.adapterPackage, "@ai-sdk/openai-compatible");
-  assert.equal(model.limits?.context, 262144);
-  assert.equal(model.limits?.output, 8192);
-  assert.equal(model.modelId, "qwen/qwen3-235b-a22b-instruct-2507-fp8");
-  assert.equal(model.recommended, false);
-  assert.equal(model.transport, "chat_completions");
-  assert.equal(model.validationStatus, "validated");
+  assert.equal(qwen.adapterPackage, "@ai-sdk/openai-compatible");
+  assert.equal(qwen.limits?.context, 262144);
+  assert.equal(qwen.limits?.output, 8192);
+  assert.equal(qwen.modelId, "qwen/qwen3-235b-a22b-instruct-2507-fp8");
+  assert.equal(qwen.recommended, false);
+  assert.equal(qwen.transport, "chat_completions");
+  assert.equal(qwen.validationStatus, "validated");
+
+  assert.equal(minimax.adapterPackage, "@ai-sdk/openai-compatible");
+  assert.equal(minimax.limits?.context, 204800);
+  assert.equal(minimax.limits?.output, 8192);
+  assert.equal(minimax.modelId, "minimaxai/minimax-m2.7");
+  assert.equal(minimax.recommended, false);
+  assert.equal(minimax.transport, "chat_completions");
+  assert.equal(minimax.validationStatus, "validated");
   assert.deepEqual(getValidatedModelKeys(), [
     "kimi-k2.6",
     "qwen3-235b-a22b-instruct-2507-fp8",
+    "minimax-m2.7",
   ]);
   assert.equal(getRecommendedProductionDefaultModel()?.key, "kimi-k2.6");
 });
 
-test("managed provider config includes the required Kilo output token limit", () => {
+test("managed provider config includes the required Kilo model limits", () => {
   const providerConfig = buildManagedProviderConfig("kimi-k2.6") as {
     models: Record<string, { limit?: { context?: number; output?: number } }>;
   };
 
   assert.equal(providerConfig.models["kimi-k2.6"]?.limit?.context, 262144);
   assert.equal(providerConfig.models["kimi-k2.6"]?.limit?.output, 8192);
+  assert.equal(providerConfig.models["minimax-m2.7"]?.limit?.context, 204800);
+  assert.equal(providerConfig.models["minimax-m2.7"]?.limit?.output, 8192);
 });
 
 test("managed provider config exposes the validated catalog for model switching", () => {
@@ -174,10 +186,16 @@ test("managed provider config exposes the validated catalog for model switching"
   assert.deepEqual(Object.keys(providerConfig.models), [
     "kimi-k2.6",
     "qwen3-235b-a22b-instruct-2507-fp8",
+    "minimax-m2.7",
   ]);
   assert.equal(providerConfig.models["kimi-k2.6"]?.id, "moonshotai/Kimi-K2.6");
   assert.equal(
     providerConfig.models["qwen3-235b-a22b-instruct-2507-fp8"]?.id,
     "qwen/qwen3-235b-a22b-instruct-2507-fp8",
   );
+  assert.equal(
+    providerConfig.models["minimax-m2.7"]?.id,
+    "minimaxai/minimax-m2.7",
+  );
+  assert.equal(providerConfig.models["minimax-m2.7"]?.name, "MiniMax M2.7");
 });