GPT-6.2 Sol Text generation API

openai / gpt-6.2-sol

An OpenAI reasoning model for implementing and reviewing code, interpreting images, and connecting evidence across long documents.

Use openai/gpt-6.2-sol through the sjolt LLM API.

Model IDopenai/gpt-6.2-sol
Input
TextImage
Output
Text
Pricing

Token rates

Input
$0.8 / 1M tokens
Cache read
$0.04 / 1M tokens
Cache write
$1 / 1M tokens
Output
$4 / 1M tokens
GPT-6.2 Solopenai/gpt-6.2-sol

Start a conversation

Write a message below to begin.

Model overview

GPT-6.2 Sol for coding, document analysis, and visual reasoning

Work from source code, documents, or images to a result you can check. The illustrations show example workflows.

GPT-6.2 Sol text generation model cover
01 · Text generation

Implement and verify software changes

Include the relevant code, a reproduction or failing check, and the conditions the result has to meet. Ask for the smallest change that passes those checks.

  • Send one required user prompt with optional system instructions.
  • Review the exact non-streaming API request before running it.
  • Inspect assistant text, finish reason, response ID, and token usage together.
GPT-6.2 Sol long-context text generation model cover
02 · Input and output

Compare findings across documents

Label each source, state the question, and ask the model to cite supporting passages and flag conflicting evidence.

Input
TextImage
Output
Text
GPT-6.2 Sol: Explain what an image shows
03 · API integration

Explain what an image shows

Pair a screenshot or technical diagram with a specific question. Ask for observable details and have the model indicate when the image leaves a point uncertain.

  • Compare enabled language models in one place.
  • Copy the request preview to keep application requests aligned with the Playground.
  • Handle assistant responses and provider availability states in the same workflow.

Model characteristics

Better for creative validation than isolated one-off generation.

01

Context and output

OpenAI specifies a 1,050,000-token context window, with up to 922,000 input tokens and 128,000 output tokens. Reasoning counts toward the output limit.

02

Reasoning effort

Choose low, medium, high, xhigh, max. The Playground starts at medium.

03

Native API workflow

Use Chat Completions for messages and Responses for application function calls. When continuing a Responses conversation, pass earlier output items, including encrypted reasoning items, back in order.

FAQ

These are the first questions to answer when evaluating this model.

Which model ID should I send?

Send openai/gpt-6.2-sol to /v1/chat/completions or /v1/responses. Both endpoints accept streaming requests.

How do I include an image?

Attach an image in the Playground, use an image_url block in Chat Completions, or use an input_image block in Responses. The model returns text.