GPT-6.2 Astra API

openai/gpt-6.2-astra
1,050,000 tokens · 1280 input / 6400 output credits / 1M tokens

The GPT-6.2 Astra API runs OpenAI's newest frontier model for professional documents, scientific analysis, and complex coding, with a 1.05M-token context and image input. Vidgo prices the GPT-6.2 Astra API at $6.40 input and $32 output per 1M tokens, 20% below OpenAI.

GPT-6.2 Astra

OpenAI · chat-completions

Chat with GPT-6.2 Astra

Each model has its own conversation. Switching never sends another model's history, and switching back resumes where you left off. Requests are billed from actual token usage.

Ctrl / ⌘ + Enter to send0 / 32,000

Continue with

GPT-6.2 Astra

GPT-6.2 Astra is the frontier model OpenAI released on November 17, 2026, replacing GPT-6 Astra for its hardest projects. Through the GPT-6.2 Astra API you send text and images and get text back, with up to 1,050,000 tokens of context and 128,000 output tokens. OpenAI reports higher scores than GPT-6 Astra on professional work, science, and coding benchmarks at a lower cost per task.

Why Choose the GPT-6.2 Astra API?

  • 20% below OpenAI pricingThe GPT-6.2 Astra API on Vidgo costs $6.40 per 1M input tokens and $32 per 1M output tokens, compared with OpenAI's $8 and $40. Cached input is $0.32 per 1M tokens.

  • Stronger than GPT-6 Astra at a lower costOpenAI reports GPT-6.2 Astra scores 5.2 points higher on Agents' Last Exam and 9 points higher on Terminal-Bench-Science than GPT-6 Astra, at less than a third of the cost per task.

  • Documents, slides, and spreadsheetsAsk for decks, workbooks, memos, and websites that follow your template. On GDPval v2 it scores 6.2 points above GPT-6 Astra with 22% fewer output tokens.

  • Complex coding in real repositoriesImplement and test multi-file changes in an existing codebase. Early testers reported cleaner code, more focused tests, and fewer rounds of revision.

  • 1.05M context, images, and toolsSend long files and screenshots in one request, call function tools through the Responses API, and set reasoning effort from low to max.

Parameters

ParameterRequirementDescription
modelRequired

Use openai/gpt-6.2-astra with either protocol.

messagesRequired for Chat Completions

A non-empty ordered array of role and content messages; content can combine text and image_url blocks. Supply system instructions in a system-role message.

inputRequired for Responses

Non-empty text or a structured message array that can combine text and image content.

instructionsOptional for Responses

System instructions that guide the response.

temperatureOptional

Sampling temperature from 0 to 2. The playground sends 1.

Default1
top_pOptional

Nucleus sampling probability from 0 to 1. The playground sends 1.

Default1
max_tokensOptional for Chat Completions

A positive integer up to 128,000 output tokens, including reasoning tokens. The playground offers 256–8,192 and starts at 2,048.

max_output_tokensOptional for Responses

A positive integer up to 128,000 output tokens, including reasoning tokens. The playground offers 256–8,192 and starts at 2,048.

reasoning.effortOptional for Responses

Sets how much the model reasons before answering.

Defaultmediumlowhighxhighmax
toolsOptional for Responses

Configure function tools available to your agent workflow.

streamOptional

Use false for complete JSON or true for SSE events. The playground streams responses.

Defaultfalsetrue

How to Use the GPT-6.2 Astra API

  1. Get a Vidgo API keyCreate an API key in the Vidgo dashboard and add credits to your account.

  2. Set the model IDSend requests to /v1/chat/completions or /v1/responses with model set to openai/gpt-6.2-astra.

  3. Add your task and filesPut instructions, documents, data, or screenshots in messages or input, and set reasoning.effort in Responses requests.

  4. Read the answer and usageEach response returns the answer and its token usage, which is what you are billed for.

GPT-6.2 Astra API Pricing

The GPT-6.2 Astra API is billed per 1M tokens from the usage in each response. Input, output, cache reads, and cache writes are counted separately, and reasoning tokens count as output.

UsageRateDetails
Input$6.40 · 1280 credits / 1M tokensStandard input tokens after reported cache reads and writes, including images.
Output$32.00 · 6400 credits / 1M tokensAnswer tokens plus reasoning tokens reported in usage.
Cache Read$0.32 · 64 credits / 1M tokensInput tokens read from cache.
Cache Write$8.00 · 1600 credits / 1M tokensInput tokens written to cache.

GPT-6.2 Astra API Use Cases

  • Financial models and client decksTurn source files into valuation workbooks, investment memos, and slides in your firm's template.

  • Legal memos and contract reviewPull the relevant facts from case files and draft memos and reviews in your team's format.

  • Scientific data analysisWrite and run analysis code on experimental data, then explain what the results show.

  • Large codebase changesPlan, implement, and test a multi-file change, then summarize what changed and how it was checked.

GPT-6.2 Astra API Tips

  • Attach your template or a finished example when you need a document in a fixed format.
  • Start with medium reasoning effort and move to xhigh or max for long research or multi-file code changes.
  • Leave room in the output limit for reasoning tokens at higher effort levels.
  • Use the Responses API for function tools and return each tool result before asking for the next step.

GPT-6.2 Astra API Notes

  • Model ID: openai/gpt-6.2-astra. Endpoints: /v1/chat/completions and /v1/responses, both with streaming.
  • Context window: 1,050,000 tokens. Maximum output: 128,000 tokens. Knowledge cutoff: May 18, 2026.
  • The playground defaults to Chat Completions and also offers Responses. Its Max Tokens control ranges from 256 to 8,192, with a default of 2,048.
  • Reasoning effort accepts low, medium, high, xhigh, and max; medium is the default.

Related Models

GPT-6.2 Astra API FAQ

What is the GPT-6.2 Astra API?

The GPT-6.2 Astra API gives developers access to GPT-6.2 Astra, the frontier model OpenAI released on November 17, 2026 for professional documents, scientific research, and complex coding. It reads text and images, returns text, and supports a 1,050,000-token context with up to 128,000 output tokens. Compared with GPT-6 Astra, OpenAI reports higher benchmark scores at a lower cost per task. You can call it programmatically or try it from the playground above.

How much does the GPT-6.2 Astra API cost?

On Vidgo, the GPT-6.2 Astra API costs $6.40 (1280 credits) per 1M input tokens and $32 (6400 credits) per 1M output tokens. Cache reads cost $0.32 and cache writes $8 per 1M tokens. That is 20% below OpenAI's $8 / $40 list price. Reasoning tokens are billed as output, and you pay only for the tokens each response uses.

Is the GPT-6.2 Astra API available now?

Yes. OpenAI released GPT-6.2 Astra on November 17, 2026, in ChatGPT, Codex, and the API. The GPT-6.2 Astra API is live on Vidgo under the model ID openai/gpt-6.2-astra, and any Vidgo API key can call it.

How do I call the GPT-6.2 Astra API?

Send a POST request to https://api.vidgo.ai/v1/chat/completions with model openai/gpt-6.2-astra and a messages array, or to /v1/responses with an input field. Pass your Vidgo API key as a Bearer token; both endpoints support streaming. The API tab on this page has ready-to-copy cURL, JavaScript, and Python examples.

How is GPT-6.2 Astra different from GPT-6 Astra?

OpenAI reports that GPT-6.2 Astra scores 5.2 points higher on Agents' Last Exam v1.1 and 9 points higher on Terminal-Bench-Science 0.1, each at less than a third of the cost per task, and 6.2 points higher on GDPval v2 with 22% fewer output tokens. The context window and output limit are the same. It is also cheaper: 1280 input and 6400 output credits per 1M tokens on Vidgo, compared with 1600 and 8000 for GPT-6 Astra.

Should I use the GPT-6.2 Astra API or the GPT-6.2 Sol API?

Use the GPT-6.2 Astra API for your hardest work: client deliverables, scientific research, and complex codebase changes. Use the GPT-6.2 Sol API for everyday coding, writing, and agents at one-quarter of Astra's token price. On FrontierCode 2.0, OpenAI reports Sol within 0.5 points of Astra at about one-fifth of the cost, both at max effort.

What is the GPT-6.2 Astra API context window?

GPT-6.2 Astra has a 1,050,000-token context window and returns up to 128,000 output tokens, including reasoning tokens. Its knowledge cutoff is May 18, 2026.

What are the GPT-6.2 Astra benchmarks?

OpenAI reports new highs for GPT-6.2 Astra on Agents' Last Exam v1.1 (5.2 points above GPT-6 Astra), Terminal-Bench-Science 0.1 (9 points above), and its internal GDPval v2 (6.2 points above, with 22% fewer output tokens). On FrontierCode 2.0, which tests merge-ready changes in real codebases, it scores above GPT-6 Astra at a lower cost.

Which reasoning effort levels does the GPT-6.2 Astra API support?

Set reasoning.effort in a Responses request to low, medium, high, xhigh, or max; medium is the default. Use medium for most tasks and xhigh or max for long research or multi-file coding, which use more output tokens.

Does the GPT-6.2 Astra API support function calling?

Yes. Define function tools in a Responses request; your application runs each tool call and sends back the result, and GPT-6.2 Astra plans the next step. It also accepts images, such as charts, slide templates, and scanned pages, as image_url blocks in Chat Completions or image content in Responses.