Add Cerebras PAUG provider and split models

Update default model provider to “Cerebras PAUG”. Reorganize the
Cerebras language‑model configuration into two groups – “Cerebras FREE”
and “Cerebras PAUG” – each with its own model list and adjusted token
limits.

modules/home/zed/user-settings/agent.nix

@@ -3,7 +3,7 @@ _: {
     always_allow_tool_actions = true;
     default_profile = "ask";
     default_model = {
-      provider = "Cerebras";
+      provider = "Cerebras PAUG";
       model = "gpt-oss-120b";
     };
   };

modules/home/zed/user-settings/language-models.nix

@@ -1,10 +1,11 @@
 _: {
-  programs.zed-editor.userSettings.language_models.openai_compatible."Cerebras" = {
+  programs.zed-editor.userSettings.language_models.openai_compatible = {
+    "Cerebras FREE" = {
       api_url = "https://api.cerebras.ai/v1";
       available_models = [
         {
-        name = "qwen-3-235b-a22b-instruct-2507";
+          name = "gpt-oss-120b";
-        display_name = "Qwen 3 235B Instruct";
+          display_name = "OpenAI GPT OSS";
           max_tokens = 65000;
           max_output_tokens = 32000;
           capabilities = {
@@ -14,6 +15,30 @@ _: {
             prompt_cache_key = false;
           };
         }
+        {
+          name = "zai-glm-4.7";
+          display_name = "Z.ai GLM 4.7";
+          max_tokens = 64000;
+          max_output_tokens = 40000;
+          capabilities = {
+            tools = true;
+            images = true;
+            parallel_tool_calls = true;
+            prompt_cache_key = true;
+          };
+        }
+        {
+          name = "llama3.1-8b";
+          display_name = "Llama 3.1 8B";
+          max_tokens = 8000;
+          max_output_tokens = 8000;
+          capabilities = {
+            tools = true;
+            images = true;
+            parallel_tool_calls = true;
+            prompt_cache_key = true;
+          };
+        }
         {
           name = "llama-3.3-70b";
           display_name = "Llama 3.3 70B";
@@ -27,10 +52,39 @@ _: {
           };
         }
         {
-        name = "gpt-oss-120b";
+          name = "qwen-3-32b";
-        display_name = "OpenAI GPT OSS";
+          display_name = "Qwen 3 32B";
+          max_tokens = 65000;
+          max_output_tokens = 8000;
+          capabilities = {
+            tools = true;
+            images = true;
+            parallel_tool_calls = false;
+            prompt_cache_key = true;
+          };
+        }
+        {
+          name = "qwen-3-235b-a22b-instruct-2507";
+          display_name = "Qwen 3 235B Instruct";
           max_tokens = 65000;
           max_output_tokens = 32000;
+          capabilities = {
+            tools = true;
+            images = true;
+            parallel_tool_calls = false;
+            prompt_cache_key = true;
+          };
+        }
+      ];
+    };
+    "Cerebras PAUG" = {
+      api_url = "https://api.cerebras.ai/v1";
+      available_models = [
+        {
+          name = "gpt-oss-120b";
+          display_name = "OpenAI GPT OSS";
+          max_tokens = 131000;
+          max_output_tokens = 40000;
           capabilities = {
             tools = true;
             images = true;
@@ -38,6 +92,67 @@ _: {
             prompt_cache_key = false;
           };
         }
+        {
+          name = "zai-glm-4.7";
+          display_name = "Z.ai GLM 4.7";
+          max_tokens = 131000;
+          max_output_tokens = 40000;
+          capabilities = {
+            tools = true;
+            images = true;
+            parallel_tool_calls = true;
+            prompt_cache_key = true;
+          };
+        }
+        {
+          name = "llama3.1-8b";
+          display_name = "Llama 3.1 8B";
+          max_tokens = 32000;
+          max_output_tokens = 8000;
+          capabilities = {
+            tools = true;
+            images = true;
+            parallel_tool_calls = true;
+            prompt_cache_key = true;
+          };
+        }
+        {
+          name = "llama-3.3-70b";
+          display_name = "Llama 3.3 70B";
+          max_tokens = 128000;
+          max_output_tokens = 65000;
+          capabilities = {
+            tools = true;
+            images = true;
+            parallel_tool_calls = true;
+            prompt_cache_key = true;
+          };
+        }
+        {
+          name = "qwen-3-32b";
+          display_name = "Qwen 3 32B";
+          max_tokens = 131000;
+          max_output_tokens = 8000;
+          capabilities = {
+            tools = true;
+            images = true;
+            parallel_tool_calls = false;
+            prompt_cache_key = true;
+          };
+        }
+        {
+          name = "qwen-3-235b-a22b-instruct-2507";
+          display_name = "Qwen 3 235B Instruct";
+          max_tokens = 131000;
+          max_output_tokens = 40000;
+          capabilities = {
+            tools = true;
+            images = true;
+            parallel_tool_calls = false;
+            prompt_cache_key = true;
+          };
+        }
       ];
     };
+  };
 }