feat: enhance models section

Author: thisisjoshford

docs/cloud/introduction.mdx

@@ -63,6 +63,8 @@ Trusted Execution Environments (TEEs) are encrypted secure areas inside CPUs and
 
 With NEAR AI Cloud, your data remains truly private. Model providers, cloud infrastructure providers, and NEAR itself cannot see, access, mine, or use your data to train models. You get the power of leading AI models — like DeepSeek, Llama, OpenAI, and Qwen — with absolute confidence in your data privacy.
 
+---
+
 ## Get Started
 
 <FeatureCardGrid>

docs/cloud/models.mdx

@@ -10,9 +10,57 @@ import DeepseekIcon from '@site/static/img/icons/models/Deepseek-logo-icon.svg';
 import GPTIcon from '@site/static/img/icons/models/GPT-logo.svg';
 import QwenIcon from '@site/static/img/icons/models/Qwen_logo.svg';
 import ZaiIcon from '@site/static/img/icons/models/zai-logo.svg';
+import CodeBlock from '@theme/CodeBlock';
 
-NEAR AI Cloud offers a curated catalog of high-performance models spanning reasoning, tool use,
-and long-context understanding. Pricing is listed per million tokens for easy comparison.
+NEAR AI Cloud provides access to leading AI models, each optimized for different use cases — from advanced reasoning and tool calling to long-context processing and multilingual tasks. All models run in secure TEE environments with transparent, pay-per-use pricing.
+
+## Quick Reference
+
+<table>
+  <thead>
+    <tr>
+      <th>Model ID</th>
+      <th>Context</th>
+      <th>Input Price</th>
+      <th>Output Price</th>
+      <th>Best For</th>
+    </tr>
+  </thead>
+  <tbody>
+    <tr>
+      <td><CodeBlock language="text">deepseek-ai/DeepSeek-V3.1</CodeBlock></td>
+      <td>128K</td>
+      <td>$1.00/M</td>
+      <td>$2.50/M</td>
+      <td>Hybrid thinking mode, tool calling, agent tasks</td>
+    </tr>
+    <tr>
+      <td><CodeBlock language="text">openai/gpt-oss-120b</CodeBlock></td>
+      <td>131K</td>
+      <td>$0.20/M</td>
+      <td>$0.60/M</td>
+      <td>Open-weight, high-reasoning, agentic workflows, configurable depth</td>
+    </tr>
+    <tr>
+      <td><CodeBlock language="text">Qwen/Qwen3-30B-A3B-Instruct-2507</CodeBlock></td>
+      <td>262K</td>
+      <td>$0.15/M</td>
+      <td>$0.45/M</td>
+      <td>Ultra-long context (262K), reasoning, instruction following, multilingual</td>
+    </tr>
+    <tr>
+      <td><CodeBlock language="text">zai-org/GLM-4.6-FP8</CodeBlock></td>
+      <td>200K</td>
+      <td>$0.75/M</td>
+      <td>$2.00/M</td>
+      <td>Agentic applications, advanced coding, tool use, refined writing</td>
+    </tr>
+  </tbody>
+</table>
+
+---
+
+## Model Details
 
 <div className="doc-model-grid">
   <div className="doc-model-card">
@@ -22,7 +70,6 @@ and long-context understanding. Pricing is listed per million tokens for easy co
       </div>
       <div>
         <h3>DeepSeek V3.1</h3>
-        <p className="doc-model-provider">deepseek-ai/DeepSeek-V3.1</p>
       </div>
     </div>
     <p>
@@ -39,6 +86,8 @@ and long-context understanding. Pricing is listed per million tokens for easy co
       <span>$1.00 /M input tokens</span>
       <span>$2.50 /M output tokens</span>
     </div>
+    <p><strong>Model ID:</strong></p>
+    <CodeBlock language="text">deepseek-ai/DeepSeek-V3.1</CodeBlock>
   </div>
 
   <div className="doc-model-card">
@@ -48,23 +97,24 @@ and long-context understanding. Pricing is listed per million tokens for easy co
       </div>
       <div>
         <h3>GPT OSS 120B</h3>
-        <p className="doc-model-provider">openai/gpt-oss-120b</p>
       </div>
     </div>
     <p>
-      GPT OSS 120B is OpenAI&rsquo;s 117B-parameter Mixture-of-Experts model for production reasoning and
-      agentic workflows. It activates just 5.1B parameters per pass, runs efficiently on a single H100
-      via native MXFP4 quantization, and supports configurable reasoning depth.
+      GPT OSS 120B is an open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI
+      designed for high-reasoning, agentic, and general-purpose production use cases. It activates 5.1B
+      parameters per forward pass and is optimized to run on a single H100 GPU with native MXFP4 quantization.
     </p>
     <p>
-      You get full chain-of-thought visibility, native tool use (function calling, browsing, structured
-      outputs), and high reliability for complex pipelines.
+      The model supports configurable reasoning depth, full chain-of-thought access, and native tool use,
+      including function calling, browsing, and structured output generation.
     </p>
     <div className="doc-model-meta">
       <span>131K context</span>
       <span>$0.20 /M input tokens</span>
       <span>$0.60 /M output tokens</span>
     </div>
+    <p><strong>Model ID:</strong></p>
+    <CodeBlock language="text">openai/gpt-oss-120b</CodeBlock>
   </div>
 
   <div className="doc-model-card">
@@ -74,23 +124,22 @@ and long-context understanding. Pricing is listed per million tokens for easy co
       </div>
       <div>
         <h3>Qwen3 30B A3B Instruct 2507</h3>
-        <p className="doc-model-provider">Qwen/Qwen3-30B-A3B-Instruct-2507</p>
       </div>
     </div>
     <p>
-      Qwen3-30B-A3B-Instruct-2507 is a 30.5B-parameter MoE model (3.3B active per inference) with an
-      ultra-long 262K context window. It excels at instruction following, logical reasoning, coding,
-      mathematics, multilingual tasks, and preference alignment&mdash;all in non-thinking mode.
-    </p>
-    <p>
-      Use it when you need multilingual comprehension and strong instruction adherence without the
-      overhead of a full reasoning model.
+      Qwen3-30B-A3B-Instruct-2507 is a mixture-of-experts (MoE) causal language model featuring 30.5 billion
+      total parameters and 3.3 billion activated parameters per inference. It supports ultra-long context up
+      to 262K tokens and operates exclusively in non-thinking mode, delivering strong enhancements in
+      instruction following, reasoning, logical comprehension, mathematics, coding, multilingual understanding,
+      and alignment with user preferences.
     </p>
     <div className="doc-model-meta">
       <span>262K context</span>
       <span>$0.15 /M input tokens</span>
       <span>$0.45 /M output tokens</span>
     </div>
+    <p><strong>Model ID:</strong></p>
+    <CodeBlock language="text">Qwen/Qwen3-30B-A3B-Instruct-2507</CodeBlock>
   </div>
 
   <div className="doc-model-card">
@@ -100,22 +149,25 @@ and long-context understanding. Pricing is listed per million tokens for easy co
       </div>
       <div>
         <h3>GLM-4.6 FP8</h3>
-        <p className="doc-model-provider">zai-org/GLM-4.6-FP8</p>
       </div>
     </div>
     <p>
-      GLM-4.6 FP8 from Zhipu AI packs 358B parameters into an FP8-quantized deployment with a 128K
-      context window. It shines in advanced coding, multi-step reasoning, and tool calling while
-      boosting token efficiency by up to 15% versus GLM-4.5.
+      GLM-4.6 is the latest flagship model in the GLM (General Language Model) series by Z.ai (formerly Zhipu AI).
+      It is oriented toward agentic applications: reasoning, tool usage, coding/engineering workflows, and long-context tasks.
+      The FP8 quantized version maintains full performance while optimizing for efficient deployment.
     </p>
     <p>
-      Positioned as a competitor to Claude Sonnet 4 and DeepSeek-V3.1-Terminus, it delivers premium
-      writing quality and robust agentic workflow support for production environments.
+      Compared with GLM-4.5, GLM-4.6 brings several key improvements: a longer 200K context window (expanded from 128K),
+      superior coding performance with better real-world results in applications like Claude Code and Cline, advanced
+      reasoning with tool use during inference, more capable search-based agents, and refined writing that better aligns
+      with human preferences in style and readability.
     </p>
     <div className="doc-model-meta">
-      <span>131K context</span>
+      <span>200K context</span>
       <span>$0.75 /M input tokens</span>
       <span>$2.00 /M output tokens</span>
     </div>
+    <p><strong>Model ID:</strong></p>
+    <CodeBlock language="text">zai-org/GLM-4.6-FP8</CodeBlock>
   </div>
 </div>

package.json

@@ -39,6 +39,6 @@
     ]
   },
   "engines": {
-    "node": ">=18.0"
+    "node": ">=20.0.0"
   }
 }