PL Log inOpen Studio
CatalogVideo36Image31Audio5Chat & code27New AI modelsCost calculatorAI agents

Chat & code OpenAI · 2026-11

GPT-6.2 Astra API

OpenAI's most capable GPT-6.2 model - for hard reasoning, coding and documents

Input$8.10per 1M tokens
Output$40.50per 1M tokens
Cached input$0.405per 1M tokens

1,000 chatbot replies$28.351,500 input + 400 output tokens each

Prompt caching works through the API: repeated context is billed at the cached rate. Failed requests cost nothing.

Compare with other chat models →

GPT-6.2 Astra is the most capable model in OpenAI's GPT-6.2 series, built for hard reasoning, coding and document work. It takes text and images, has a 1,050,000-token context window (up to 922,000 input tokens) and returns up to 128,000 tokens. You set the reasoning depth with reasoning_effort: low, medium (default), high or xhigh. You can force a JSON schema with response_format, and streaming works too.

max_completion_tokens includes reasoning tokens, and those bill as output - with a very low cap you may get only reasoning and no visible answer.

Best for

  • The hardest engineering tasks and reviews of large code changes
  • Long-context analysis of documents, contracts and reports
  • Strict JSON output for automation
  • Final verification of answers from cheaper models

How to call it

OpenAI SDK with a different base URL:

python
from openai import OpenAI

client = OpenAI(base_url="https://jantaris.com/v1", api_key="sk-jt-...")
resp = client.chat.completions.create(
    model="gpt-6.2-astra",
    reasoning_effort="medium",
    messages=[{"role": "user", "content": "Review this database migration plan and list the risks."}],
)
print(resp.choices[0].message.content)

FAQ

What is the GPT-6.2 Astra API price?

$8.10 / $40.50 per 1M tokens. Prompts over 272K input tokens are billed at the higher long-context rate for the whole request.

Astra or Sol?

GPT-6.2 Sol is the cheaper model of the same series for everyday coding and agents. A simple split: most traffic goes to Sol, and only what Sol cannot handle goes to Astra. Both run on one key and one balance.

Does tool calling work?

Not on this endpoint - OpenAI offers tools for the GPT-6.2 series only through the Responses API. For agent loops with tools use GPT-5.6 Sol.

Does it accept images?

Yes, as image_url parts in a message.

How to call GPT-6.2 Astra via API

from openai import OpenAI

client = OpenAI(base_url="https://jantaris.com/v1", api_key="sk-jt-...")
resp = client.chat.completions.create(
    model="gpt-6.2-astra",
    messages=[{"role": "user", "content": "Write a haiku about Baltic amber."}],
)
print(resp.choices[0].message.content)