From 0017c96e7b802fa4287f3761d7510a03d562fb82 Mon Sep 17 00:00:00 2001
From: Guillaume Gay <guillaume.gay@protonmail.com>
Date: Mon, 20 Jul 2026 04:32:59 +0000
Subject: [PATCH] feat: add alibaba/glm-5.2-fast-preview

---
 models/alibaba/glm-5.2-fast-preview.yaml      | 44 ++++++++++++++
 packages/modelparams/src/generated/data.ts    | 60 +++++++++++++++++++
 .../modelparams/src/generated/defaults.ts     |  4 ++
 .../modelparams/src/generated/model-ids.ts    |  1 +
 .../modelparams/src/generated/params-by-id.ts |  7 +++
 5 files changed, 116 insertions(+)
 create mode 100644 models/alibaba/glm-5.2-fast-preview.yaml

diff --git a/models/alibaba/glm-5.2-fast-preview.yaml b/models/alibaba/glm-5.2-fast-preview.yaml
new file mode 100644
index 0000000..d3c213d
--- /dev/null
+++ b/models/alibaba/glm-5.2-fast-preview.yaml
@@ -0,0 +1,44 @@
+# yaml-language-server: $schema=https://modelparams.dev/api/v1/schema.json
+provider: alibaba
+authType: api_key
+model: glm-5.2-fast-preview
+params:
+  - path: max_tokens
+    type: integer
+    label: Max tokens
+    description: Maximum number of output tokens the model may generate.
+    range:
+      min: 1
+    group: generation_length
+  - path: temperature
+    type: number
+    label: Temperature
+    description: Controls randomness. Lower values make outputs more focused; higher values make them more varied.
+    range:
+      min: 0
+      max: 2
+      step: 0.1
+    group: sampling
+  - path: top_p
+    type: number
+    label: Top P
+    description: Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability.
+    range:
+      min: 0
+      max: 1
+      step: 0.01
+    group: sampling
+  - path: extra_body.top_k
+    type: integer
+    label: Top K
+    description: Limits generation to the selected number of highest-probability tokens.
+    default: 20
+    range:
+      min: 1
+    group: sampling
+  - path: extra_body.chat_template_kwargs.enable_thinking
+    type: boolean
+    label: Enable thinking
+    description: Controls Qwen3 thinking mode when using OpenAI-compatible clients that pass provider-specific extra body fields.
+    default: true
+    group: reasoning
diff --git a/packages/modelparams/src/generated/data.ts b/packages/modelparams/src/generated/data.ts
index 96c83e5..6308a91 100644
--- a/packages/modelparams/src/generated/data.ts
+++ b/packages/modelparams/src/generated/data.ts
@@ -4,6 +4,66 @@
 import type { ModelId } from "./model-ids.js";
 
 export const CATALOG = [
+  {
+    "provider": "alibaba",
+    "authType": "api_key",
+    "model": "glm-5.2-fast-preview",
+    "params": [
+      {
+        "path": "max_tokens",
+        "label": "Max tokens",
+        "description": "Maximum number of output tokens the model may generate.",
+        "group": "generation_length",
+        "type": "integer",
+        "range": {
+          "min": 1
+        }
+      },
+      {
+        "path": "temperature",
+        "label": "Temperature",
+        "description": "Controls randomness. Lower values make outputs more focused; higher values make them more varied.",
+        "group": "sampling",
+        "type": "number",
+        "range": {
+          "min": 0,
+          "max": 2,
+          "step": 0.1
+        }
+      },
+      {
+        "path": "top_p",
+        "label": "Top P",
+        "description": "Controls nucleus sampling by limiting generation to tokens within the selected cumulative probability.",
+        "group": "sampling",
+        "type": "number",
+        "range": {
+          "min": 0,
+          "max": 1,
+          "step": 0.01
+        }
+      },
+      {
+        "path": "extra_body.top_k",
+        "label": "Top K",
+        "description": "Limits generation to the selected number of highest-probability tokens.",
+        "group": "sampling",
+        "type": "integer",
+        "default": 20,
+        "range": {
+          "min": 1
+        }
+      },
+      {
+        "path": "extra_body.chat_template_kwargs.enable_thinking",
+        "label": "Enable thinking",
+        "description": "Controls Qwen3 thinking mode when using OpenAI-compatible clients that pass provider-specific extra body fields.",
+        "group": "reasoning",
+        "type": "boolean",
+        "default": true
+      }
+    ]
+  },
   {
     "provider": "alibaba",
     "authType": "api_key",
diff --git a/packages/modelparams/src/generated/defaults.ts b/packages/modelparams/src/generated/defaults.ts
index a3bb2df..70e5d05 100644
--- a/packages/modelparams/src/generated/defaults.ts
+++ b/packages/modelparams/src/generated/defaults.ts
@@ -5,6 +5,10 @@ import type { ModelId } from "./model-ids.js";
 import type { ParamsById } from "./params-by-id.js";
 
 export const DEFAULTS = {
+  "alibaba/glm-5.2-fast-preview": {
+    "extra_body.top_k": 20,
+    "extra_body.chat_template_kwargs.enable_thinking": true,
+  },
   "alibaba/qwen-flash": {
     "extra_body.top_k": 20,
     "extra_body.chat_template_kwargs.enable_thinking": true,
diff --git a/packages/modelparams/src/generated/model-ids.ts b/packages/modelparams/src/generated/model-ids.ts
index 99530ee..865a8e5 100644
--- a/packages/modelparams/src/generated/model-ids.ts
+++ b/packages/modelparams/src/generated/model-ids.ts
@@ -2,6 +2,7 @@
 // Source of truth: the YAML catalog under /models in modelparams.dev.
 
 export const MODEL_IDS = [
+  "alibaba/glm-5.2-fast-preview",
   "alibaba/qwen-flash",
   "alibaba/qwen-plus",
   "alibaba/qwen3-coder-flash",
diff --git a/packages/modelparams/src/generated/params-by-id.ts b/packages/modelparams/src/generated/params-by-id.ts
index 651d6d4..1e7bb8c 100644
--- a/packages/modelparams/src/generated/params-by-id.ts
+++ b/packages/modelparams/src/generated/params-by-id.ts
@@ -7,6 +7,13 @@
  * parameters they want to override.
  */
 export type ParamsById = {
+  "alibaba/glm-5.2-fast-preview": {
+    max_tokens: number;
+    temperature: number;
+    top_p: number;
+    "extra_body.top_k": number;
+    "extra_body.chat_template_kwargs.enable_thinking": boolean;
+  };
   "alibaba/qwen-flash": {
     max_tokens: number;
     temperature: number;