OpenAI’s frontier reasoning model for advanced coding, research, computer use, cybersecurity, design, and long-running agentic workflows.
Frontier reasoning model
OpenAI’s highest-capability GPT-5.6 model for complex professional workflows
1M+ token context
Supports a 1,050,000-token context window for large codebases and extensive documents
Large output capacity
Generates up to 128,000 output tokens in a single response
Advanced coding
Built for terminal workflows, repository-scale changes, debugging, and multi-step software engineering
Agentic workflows
Plans, executes, verifies, and coordinates long-running tasks using tools
Multimodal understanding
Accepts both text and image inputs
Computer use
Can interact with graphical interfaces and complete browser or desktop-based tasks
Extensive tool support
Supports web search, file search, code interpreter, hosted shell, computer use, MCP, and image generation
Configurable reasoning
Supports multiple reasoning levels for balancing speed, depth, and cost
Strong design judgment
Improved layout, visual hierarchy, frontend implementation, and design-system awareness
Cybersecurity capabilities
Designed for vulnerability research, defensive testing, debugging, and patch development
API and ChatGPT access
Available through the OpenAI API, ChatGPT, Codex, and supported enterprise workflows
- Build and refactor applications across large, multi-file codebases
- Debug complex software issues using terminal and repository tools
- Create production-ready frontend interfaces from detailed requirements
- Review architecture and plan large-scale code migrations
- Analyze long technical documents, specifications, and research papers
- Build autonomous and semi-autonomous AI agents
- Perform defensive security reviews and vulnerability analysis
- Generate patches and remediation plans for software vulnerabilities
- Automate browser and computer-based professional workflows
- Analyze screenshots, diagrams, interfaces, and other visual inputs
- Conduct scientific, technical, and knowledge-work research
- Coordinate multi-step workflows involving files, APIs, code, and external tools
Available through the OpenAI API with usage-based token pricing.
Standard API pricing per 1 million tokens:
Input: $5.00
Cached input: $0.50
Output: $30.00
Requests containing more than 272K input tokens are charged at 2× the standard input rate and 1.5× the standard output rate for the full request.
Cache writes are billed at 1.25× the uncached input-token rate.
GPT-5.6 Sol is also available on eligible paid ChatGPT plans. Availability and usage limits vary by plan, workspace settings, region, and product.
Commercial usage is subject to OpenAI’s applicable service terms, usage policies, and API terms.
Integration highlights
- OpenAI Responses API
- OpenAI Chat Completions API
- ChatGPT
- Codex
- OpenAI Batch API
- OpenAI File Search
- OpenAI Web Search
- OpenAI Code Interpreter
- Hosted Shell
- Computer Use
- Model Context Protocol (MCP)
- Image Generation
- Amazon Bedrock