Skip to content

Model detail

GPT-6.2 Astra

Provided by OpenAI
Pay-as-you-go Dynamic pricing 2 tiers

Reasoning model for complex software and document projects.

Model specs

Model ID
gpt-6.2-astra
Context length
1.05M
Max output
128K
I/O modalities
Released
2026-11

02

API endpoints

One MixRoute gateway, OpenAI-compatible

  • OpenAI-compatible /v1/chat/completions POST

03

Tiered pricing

Unit: /1M Tokens

Pricing updated

Tier Input /1M Tokens Output /1M Tokens Cache read /1M Tokens Cache write /1M Tokens Cache write (12h) /1M Tokens
standard Length ≤ 272K $8.0000 $40.0000 $0.4000 $10.0000 $20.0000
long_context Length > 272K $16.0000 $60.0000 $0.8000 $20.0000 $40.0000

05

Selection summary

Quickly judge whether it fits your workload

What is this model good for?

Reasoning

GPT-6.2 Astra is built for multi-step reasoning, and it reads images alongside text. It accepts text and image input and returns text. Its capabilities are broken down in the section just below. Give it one problem from your own workload and check the reasoning trace before you commit.

What should you check before using it?

Pricing changes at 272,000 input tokens

GPT-6.2 Astra bills inputs above 272,000 tokens at 16.00 USD input and 60.00 USD output per 1M tokens, against 8.00 USD and 40.00 USD below that line. Size a representative task of yours to estimate which tier your work falls in.

Why use it through MixRoute?

One MixRoute key reaches GPT-6.2 Astra

One MixRoute key covers gpt-6.2-astra, on the same account as the rest of the catalog. MixRoute supports calling GPT-6.2 Astra through its OpenAI-compatible endpoint; point your client at https://api.mixroute.ai/v1 with the model ID gpt-6.2-astra and re-run your integration tests.

07

Capabilities & compatibility

Unconfirmed capabilities stay "Unknown," no guesses

Available:

Tool calling

GPT-6.2 Astra supports function and tool calls.

EX Send a tools definition; the response returns tool_calls to execute and feed results back.
Available:

Structured output

GPT-6.2 Astra can produce JSON-formatted output on demand.

EX response_format: { type: "json_schema" } forces valid JSON output.
Available:

Streaming

GPT-6.2 Astra supports streamed responses.

EX stream: true streams tokens incrementally, so chat UIs render as they arrive.
Available:

Prompt caching

GPT-6.2 Astra supports prompt caching under supported routing conditions.

EX Keep the system prompt and tool definitions at the start of the request; repeated prefixes are automatically billed at the cache-read rate.
Available:

Vision input

GPT-6.2 Astra can process image inputs alongside text.

EX Pass a base64-encoded image in a image_url content part alongside text messages.

08

Token cost estimator

Live estimate from this page's pricing, not an actual bill

Share of the same prefix read repeatedly, up to 100%

09

FAQ

GPT-6.2 Astra is a reasoning model. Reasoning effort can be tuned per request. Lower effort settings cut reasoning tokens and cost for routine work; reserve higher settings for problems that need deeper search. Measure quality at two or three settings before fixing one in production.

One endpoint, a testable decision

Test this model and alternate routes with the same request format

Start from a real workload, then let quality, total cost, and failure conditions decide whether to send production traffic.