Announcement: OpenAI GPT 5.6 Sol Pricing Adjustment

Important Notice: OpenAI has implemented a temporary price reduction for the GPT 5.6 Sol model. This promotional pricing is guaranteed to remain in effect until at least November 21.

For a comprehensive index of all documentation, please refer to the llms.txt file. Users can access Markdown versions of any documentation page by simply appending .md to the end of the URL.

---

🛠️ API Ecosystem & Navigation

The OpenAI platform is divided into several core functional areas. Below is the architectural breakdown of the current documentation suite.

1. Core Model Capabilities

The API provides a wide array of modalities and reasoning capabilities:

• Text & Code:
  • Standard text generation and specialized code generation.
  • Implementation of structured output for predictable data.
• Reasoning:
  • Advanced reasoning models with adjustable reasoning_effort.
  • Best practices for complex logical chains.
• Multimodal:
  • Vision: Image inputs and analysis.
  • Media: Image and video generation.
  • Audio: Realtime speech, transcription, and voice agents (supporting WebRTC, WebSocket, and SIP).

2. Agentic Frameworks

The transition from simple prompts to autonomous agents is managed via:

• Agents SDK: Defining agents, running them in sandboxes, and managing orchestration.
• Assistants API: Deep dives into tool integration and model selection.
• ChatKit: Customizing widgets and advanced integrations.

3. Tooling & Connectivity

OpenAI utilizes a robust set of connectors to interact with the external world:

---

📉 Pricing & Cost Analysis

The current cost reduction for GPT 5.6 Sol can be represented by the following cost function:

$\text{Total Cost} = (\text{Input Tokens} \times \text{Reduced Rate}) + (\text{Output Tokens} \times \text{Reduced Rate})$

Price Change Summary:

• ~~Standard Rate: $X.XX per 1M tokens~~$\rightarrow$**Discounted Rate:$Y.YY per 1M tokens**
• Validity: Active through Nov 21.

Cost Optimization Checklist

• Enable prompt caching to reduce redundant input costs.
• Utilize Batch processing for non-urgent tasks.
• Implement Flex processing for throughput balancing.
• Monitor spend limits via the Admin API.

---

🏗️ Infrastructure & Governance

Deployment Workflow

The following diagram illustrates the path from development to production:

Technical Specifications

For those managing infrastructure, the following resources are available:

• Access Control: Service accounts, Project-based access, and Workload identity federation.
• Network: Private Link, IP allowlists, and IP egress ranges.
• Cloud Providers: Support for AWS, Microsoft Azure, Google Cloud, and Oracle Cloud Infrastructure.

# Example: Checking rate limits via CLI
openai api rate-limits --model gpt-5.6-sol

---

🔌 Plugins, MCP, and Extensions

The ecosystem extends beyond the primary API into specialized plugins and the Model Context Protocol (MCP).

Development Lifecycle:

1. Brainstorm: Define use cases.
2. Build: Create an MCP server $\rightarrow$ Add UI $\rightarrow$ Authenticate users.
3. Package: Build skills and package the plugin.
4. Publish: Test $\rightarrow$ Submit $\rightarrow$ Review.

Note: Documentation includes specific conversion specs for Restaurant Reservations, Quotes, and Product Checkouts, as well as guides for Claude Code plugins.

---

📢 Commercial & Workspace Integration

OpenAI provides specialized tools for business operations:

• Workspace Agents: Triggering runs via access tokens.
• Commerce API:
  • Products: Feeds, Promotions, and Ads.
  • Measurement: Pixel implementation and Conversion APIs.
  • Advertiser API: Campaign targeting and custom audiences.

Final Reference Table: Feature Maturity