GPT-6.2 Astra Text generation API

openai / gpt-6.2-astra

An OpenAI reasoning model for complex software changes, evidence synthesis across documents, and detailed visual analysis.

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

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

Token rates

Input
$2.4 / 1M tokens
Cache read
$0.12 / 1M tokens
Cache write
$3 / 1M tokens
Output
$12 / 1M tokens
GPT-6.2 Astraopenai/gpt-6.2-astra

Start a conversation

Write a message below to begin.

Model overview

GPT-6.2 Astra for complex coding and evidence review

Use the model to work through dependencies, conflicting sources, and visual evidence. The illustrations depict example workflows.

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

Trace changes across a codebase

Provide the relevant modules and a failing test or observed behavior. Ask for the smallest change that preserves surrounding contracts and for a check that verifies it.

  • 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 Astra long-context text generation model cover
02 · Input and output

Reconcile evidence across documents

Label each source and state the decision you need to make. Ask for a comparison with citations and a list of unresolved conflicts.

Input
TextImage
Output
Text
GPT-6.2 Astra: Inspect visual details
03 · API integration

Inspect visual details

Attach a diagram or screenshot and identify what the analysis should cover. Ask the model to separate the visible details from its inferences.

  • 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. Available context can vary by deployment. 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, include the earlier output items, including encrypted reasoning items.

FAQ

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

Which model ID should I send?

Send openai/gpt-6.2-astra to /v1/chat/completions or /v1/responses. Both endpoints support streaming.

How do I include an image?

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