AI KNOWLEDGE DESK

Models · entities · concepts · comparisons · practical tools

GETLLMS.ORG

🚀Function Overview

Evaluate Astra on changes that span multiple systems or require careful verification. Compare task success, retries and review time against GPT-6.2 Sol before making it your default.

Key Features

  • The OpenAI API has a 1,050,000-token context window, with up to 922,000 input tokens and 128,000 output tokens. Knowledge cutoff: May 18, 2026.
  • Text and image input, with text output. Image generation through a Responses tool is a separate capability.
  • Standard OpenAI API rates per million tokens: $8 input, $0.40 cached input, $10 cache write (30 minutes), $40 output. A 12-hour cache write costs $20 per million tokens at the short-context Standard rate.
  • Above 272,000 input tokens, the full request uses these Standard API rates per million tokens: $16 input, $0.80 cached input, $20 cache write (30 minutes), $60 output.
  • API Fast costs 2× Standard for the same context length. API Ultrafast costs 3× Standard through February 28, 2027, rising to 6× on March 1, Pacific time. Codex and ChatGPT Work have their own plan and usage rules.
  • Use Responses for tools such as web search, file search or computer use. Chat Completions supports this model for supported calls without tools.
  • OpenAI’s rollout covers eligible Codex and ChatGPT Work accounts. Check your workspace settings for access. The rollout does not include regular Chat.
  • OpenAI reports that Astra scored 5.2 points above GPT-6 Astra on Agents’ Last Exam. That result describes the published long-horizon test and its configuration.

Use Cases

  • •Trace a regression across services and verify the fix against integration tests.
  • •Build an interactive prototype and check its rendered behavior in a browser.
  • •Reconcile source documents and produce a reviewable deliverable.

⚙️Input Parameters

model

string

Set this to gpt-6.2-astra.

input

string or array

Use a text prompt or text and image content blocks in a Responses API request.

reasoning.effort

string

OpenAI suggests medium for a new integration and low when migrating from none or minimal.

tools

array

Optional tools for the Responses API. Enable only what the task requires.

💡Usage Examples

Example 1

Input Parameters

{
  "model": "gpt-6.2-astra",
  "input": "Review the supplied integration test failures. Identify the likely shared cause, list the evidence for it, and propose checks that would distinguish it from other explanations.",
  "reasoning": {
    "effort": "medium"
  }
}

Output Results

Illustrative request. Inspect the returned response and token usage when you run it with your API account.