Coding with visual context
Discuss code alongside a screenshot, diagram or reference image.
Back to modelsIntegration guide
Text output · Image inputProvider: OpenAI
GPT-5.6 Sol
gpt-5.6-solText and image input for chat and coding, with streaming, tool calling and configurable reasoning.
- Modality
- Text output · Image input
- List price (USD)
- Input 4.000000 · Output 20.000000 / million tokensCache read 0.400000 / million tokens
- Context window
- 1.1M
Model specifications and parameters
Fields are shown for this model's connected endpoint, not one generic form for every model.
| Parameter / field | Current range | Description |
|---|---|---|
| messages | Required | Submit ordered conversation messages with their roles. |
| max_completion_tokens | 1–8,192 (web test) | The web test uses a lower cost-estimate guardrail; the provider limit appears above. |
| temperature | 0–2 | Sampling temperature: higher is more varied, lower is more focused. |
| top_p | 0–1 | Nucleus sampling; V4.1 Flash's effective range changes with thinking mode. |
| stream | true / false | Return chunks as generation progresses. |
| thinking.type | Model-specific | Support and value mapping vary by model version; the web experience shows only verified controls. |
| reasoning_effort | Model-specific | Reasoning levels available for this version; adjacent levels may map to the same effective setting. |
Provider limits and LINK42 web-test limits may differ. LINK42's current quote and actual settlement govern price; external docs are provided to verify model capabilities.
Test a result before integrating
Adjust parameters, review the estimate and validate your idea with an actual output.
Test parameters
gpt-5.6-solChecking your session…
Output
Your first result starts here
Enter a prompt and start a test. Generated text, images or video will appear here.
Test history
Private to your account, including recent results and actual charges.
Explore the model and adjust parameters first. Sign in to restore your draft and run a paid test.
Overview
Input and responses
Ask a question in text, attach images, or combine both. API clients can include images in the supported request format.
Reasoning controls
Use reasoning.effort with Responses or reasoning_effort with Chat Completions. Supported levels: none, low, medium, high, xhigh and max. Set none when you do not want a reasoning step.
Use cases
Application workflows
Stream answers into your interface and decide how your application should handle each requested tool call.
Frequently asked questions
Can I include an image?
Yes. In the Experience Center, use the + button or paste an image into the message box. For API calls, follow the image-input format for your chosen endpoint.
Where do tool calls run?
In your application or agent. The model requests a tool call; your client decides whether and how to execute it.
Pricing
These are standard model prices. Signed-in estimates account for your billing identity and parameters; final charges use actual usage.
- Input price (USD)
- 4.000000/ million tokens
- Output price (USD)
- 20.000000/ million tokens
- Cache reads (USD)
- 0.400000/ million tokens
API
After validating the result, use the same model identifier in your application. Keep API keys on your server.
cURL
curl https://link42.ai/v1/chat/completions -H "Authorization: Bearer $LINK42_API_KEY" -H "Content-Type: application/json" -d '{"model": "gpt-5.6-sol", "messages": [{"role": "user", "content": "…"}]}'